E L E M E N T S OF SPACE T E C H N O L O G Y FOR AEROSPACE ENGINEERS
E L E M E N T S OF SPACE T E C H N O L O G Y FOR ...
413 downloads
2374 Views
17MB Size
Report
This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!
Report copyright / DMCA form
E L E M E N T S OF SPACE T E C H N O L O G Y FOR AEROSPACE ENGINEERS
E L E M E N T S OF SPACE T E C H N O L O G Y FOR AEROSPACE ENGINEERS
a This Page Intentionally Left Blank
ELEMENTS OF SPACE TECHNOLOGY FOR AEROSPACE E N G I N E E R S
RUDOLF X. MEYER Department of Mechanical and Aerospace Engineering University of California Los Angeles, California
ACADEMIC PRESS San Diego London Boston New York Sydney Tokyo Toronto
This book is printed on acid-free paper. | Copyright 9 1999 by Academic Press All rights reserved. No part of this publication may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopy, recording, or any information storage and retrieval system, without permission in writing from the publisher. ACADEMIC PRESS 525 B. St. Suite 1900, San Diego, California 92101-4495, USA http://www.apnet.com Academic Press 24-28 Oval Road, London NW1 7DX, UK http://www.hbuk, co.uk/ap /
Library of Congress Cataloging-in-Publication Number Meyer, Rudolf X. Elements of space technology for aerospace engineers / Rudolf X. Meyer. p. cm. Includes bibliographical references and index. ISBN 0-12-492940-0 1. Space vehiclesmDesign and construction. 2. Aerospace engineering. 3. Astronautics. I. Title. II. Title: Elements of space technology TL795 .M48 1999 629.4mddc21 98-52665 CIP Printed in the United States of America 99 00 01 02 03 IP 9 8 7 6 5 4 3 2 1
CONTENTS ix
Preface CHAPTER 1. REFERENCE FRAMES AND TIME 1.1 1.2 1.3" 1.4 1.5 CHAPTER 2.
FORCES AND MOMENTS 2.1 2.2 2.3 2.4* 2.5* 2.6
CHAPTER 3.
Gravity Thrust Aerodynamic Forces and Moments Free Molecule Flow Solar Radiation Pressure Atmospheric Entry
ORBITS AND TRAJECTORIES IN AN INVERSE SQUARE FIELD 3.1 3.2 3.3* 3.4 3.5* 3.6* 3.7 3.8* 3.9* 3.10"
CHAPTER 4.
Reference Frames Motion in Accelerated Reference Frames Example: The Yo-Yo Despin Mechanism Euler Angles and Transformations of Coordinates Time Intervals and Epoch
Kepler Orbits and Trajectories Position as a Function of Time D'Alembert and Fourier-Bessel Series Orbital Elements Spacecraft Visibility above the Horizon Satellite Observations and the f and g Series Special Orbits Perturbations by Other Astronomical Bodies Planetary Flyby and Gravity Assist Relativistic Effects
CHEMICAL ROCKET PROPULSION 4.1 4.2
Configurations of Liquid-Propellant Chemical Rocket Motors Configurations of Solid-Propellant Motors
1
7 9 12 14 21 21 31 37 40 47 49
59 60 65 66 68 71 73 75 81 85 91 97
99 i01
vi
Contents
4.3* 4.4 4.5 4.6 4.7* 4.8 4.9* 4.10" 4.11 4.12" 4.13" 4.14 4.15" 4.16" 4.17" 4.18 4.19 4.20 CHAPTER 5.
CHAPTER 6.
103 111 116 117 122 126 127 135 139 143 147 155 158 160 162 165 170 175
ORBITAL MANEUVERS
181
5.1 5.2* 5.3 5.4 5.5* 5.6 5.7 5.8* 5.9* 5.10" 5.11"
181 184 188 192 194 196 197 199 202 205 209
Minimum Energy Paths Lambert's Theorem Maneuvers with Impulsive Thrust H o h m a n n Transfers Other Transfer Trajectories On-Orbit Drift Launch Windows Injection Errors and Their Corrections On-Orbit Phase Changes Rendezvous Maneuvers Gravity Turn
ATTITUDE CONTROL 6.1 6.2* 6.3 6.4 6.5 6.6 6.7*
CHAPTER 7.
Rocket Stages Idealized Model of Chemical Rocket Motors Ideal Thrust Rocket Motor Operation in the Atmosphere Two- and Three-Dimensional Effects Critique of the Ideal Model Elements of Chemical Kinetics Chemical Kinetics Applications to Rocket Motors Liquid Propellants Propellant Tanks Propellant Feed Systems of Launch Vehicles Thrust Chambers of Liquid-Propellant Motors Pogo Instability and Prevention Thrust Vector Control Engine Control and Operations Liquid-Propellant Motors and Thrusters on Spacecraft Components of Solid-Propellant Rocket Motors Hybrid-Propellant Rocket Motors
Principal Axes and Moments of Inertia of Spacecraft The Euler Equations for Time-Dependent Moments of Inertia The Torque-Free Spinning Body Attitude Control Sensors Attitude Control Actuators Spin-Stabilized Vehicles Gravity Gradient Stabilization
SPACECRAFT THERMAL DESIGN 7.1" Fundamentals of Thermal Radiation 7.2 Spacecraft Surface Materials
215 218 223 225 230 240 251 262 269 269 278
Contents
7.3 7.4 7.5* 7.6* 7.7* 7.8 7.9
Model of a Spacecraft as an Isothermal Sphere Earth Thermal Radiation and Albedo Diurnal and Annual Variations of Solar Heating Thermal Blankets Thermal Conduction Lumped Parameter Model of a Spacecraft Thermal Control Devices
vii 281 284 285 286 289 290 299
Appendices
307
Index
325
a This Page Intentionally Left Blank
PREFACE This book is intended as a first introduction to space technology for aerospace engineering students, at either the senior or graduate level. It is hoped also to be useful to professional engineers who wish to become more familiar with the aerospace aspects of space systems. It would have been tempting to include all major disciplines that are represented in space systems engineering, including electronics, space communications, and even the civil engineering of launch site construction. That such a broad, interdisciplinary approach would have merit is suggested by the fact that quite often the effort (and cost) in developing a new spacecraft and its ground support is about equally divided between the mechanical and electronic systems. However, the division of most engineering curricula into separate aerospace and electronic engineering disciplines would have made such a collection quite impractical as a textbook. The concentration here is therefore on the mechanical ("aerospace") aspects, omitting the equally important electronic and communications disciplines. Nevertheless, in instances in which the disciplines clearly overlap, as in the thermal control of electronic components, an effort has been made to broaden the scope sufficiently to make the student aware of the problems and solutions that are addressed by the electronics and communications engineers. Most of the material in this book can be covered by graduate students in one semester and similarly by seniors at American universities. Several parts, particularly in Chapter 4 on the chemistry of propellants are suited for reading by the student, without need for class presentation. Major parts have been taught by the author to engineering students at the University of California, Los Angeles. The book has been arranged such that, at the choice of the instructor, entire sections can be omitted without losing continuity. These sections, which sometimes contain more advanced topics, are indicated in the Table of Contents by an asterisk. The problems at the end of each chapter that relate to these sections are indicated in the same manner. Chapters 1 and 2 are preliminary to the discussion of orbits and trajectories in Chapters 3 and 5. Some of this material may already be familiar to the reader but may be used for review or may be skipped. Classical orbital mechanics is treated relatively briefly, forgoing the elegance gained by Hamiltonian mechanics. The analysis of spacecraft orbits, in good part borrowed from classical celestial mechanics, has become the specialty of ix
x
Preface
numerical analysts. Instead, more importance in this book has been placed on the trajectories encountered in rendezvous and docking maneuvers and on launch trajectories with thrust and aerodynamic drag. For lack of space in the typical aerospace curriculum, more specialized topics, such as spacecraft testing methods or reliability analyses, had to be omitted or are mentioned only very briefly. Similarly, topics that are usually dealt with elsewhere in the curriculum are introduced only in instances in which their application to space technology is unique in some sense. In spite of their importance, the design and analysis of launch vehicle and spacecraft structures are therefore hardly touched upon in this book because the principles involved are very similar to those of aircraft structures, a topic that usually precedes this in the aerospace curriculum. Similarly, in the chapter on attitude control, the emphasis is on sensors and actuators, not on control theory, in which specialized courses are offered at most institutions. The author may also have to apologize for not having included a chapter on space mission design. The excuse is that at present mission designs and their future capabilities change so rapidly that it seemed better to limit the lecture course to the fundamentals that are likely to have more permanent validity. Because the aim is to emphasize the more fundamental aspects, the treatment must necessarily be in good part theoretical. However, the author hopes that he has resisted successfully the temptation to make the mathematical arguments appear to be more sophisticated than warranted by the subject matter. Where it seemed desirable, some brief mathematical comments are made. In spite of this emphasis on the scientific and engineering foundations and on analysis, an effort has been made to include the more practical aspects encountered in the design of space vehicles and their components whenever these topics were felt to be more than ephemeral, soon to be replaced by more advanced designs. Emphasis has also been placed on providing a "feel for numbers" so that the reader may develop a sense of what is numerically important and what is not. Each chapter is followed by some exercises, varying from the simple to the more difficult. It is the author's contention, however, that such exercises, which can be stated in a few lines, are less valuable to the engineering student than assigned projects, be they mission analyses or hardware designs. To support such projects, space was found to include at least some of the more important technical data in an appendix. Some technicalities: Throughout this text SI units are used. Only in a very few instances, and only when sanctioned by almost universal usage, will other units, such as electron volts for energy, appear. ESA and now also NASA publications have been using exclusively the SI units, with only an occasional and parenthetical mention of English units when dictated by very long usage, and as still occur in major data collections. For the convenience of the reader, the principal symbols used are summarized at the end of each chapter, often with the n u m b e r of the defining equation.
Preface
xi
Many of the subscripts are used uniformly throughout this book. They are, with their meaning:
( )a ( )av
ambient average ( )el electric ( )ex nozzle exit ( )g earth; also gravitational ( )c Greenwich
( )h ( )pl ( )pr ( )s ( )si
sun; solar planet propellant spacecraft sidereal
To help distinguish the forest from the trees, equation n u m b e r s that are printed in bold font indicate that the equation is either important, likely to be referred to in a later chapter, or else the final result of a preceding development. Simple corollaries or special cases of a preceding equation are given the same number, but with a prime. As would seem appropriate for a textbook of this type, journal papers or monographs are referenced only w h e n the text or illustrations are so close to the original that it would be unprofessional not to refer to the author. Instead, references are mainly to books, monographs, or data collections, where the reader m a y find additional material or different presentations. It remains for me to thank m y colleagues and friends, far too m a n y to m e n t i o n by n a m e here, at The Aerospace Corporation and at the University of California who have inspired m u c h of this work. Space technology is a national and international enterprise with m a n y workers. To t h e m go my thanks.
Ad astra pontem fecerunt fabri, in majorem Dei gloriam.
a This Page Intentionally Left Blank
'11 Reference Frames and Time This chapter includes a discussion of the reference frames commonly used in space vehicle dynamics. They are employed to describe and analyze orbits and trajectories and also to describe the orientation in space ("attitude") of the vehicle. Euler angles are introduced to characterize the rotation of reference frames. Because of the importance of relative motion in such problems as fluid sloshing in the vehicle's propellant tanks or the effect of vehicle motion on gyroscopes, the transformation equations for velocities and accelerations are rederived. Although the reader will already be familiar with this topic and other parts of this chapter, they are included because later chapters will refer to them. A second purpose is to introduce notation that will remain uniform for the remainder of this book. The chapter concludes with a discussion of the various time measurem e n t systems that are in use in the design of space missions.
1.1
Reference Frames A quantitative description of the motion of space vehicles, be they launch vehicles, upper stage vehicles, artificial satellites, or deep-space probes, requires one or several referenceframes with respect to which the vehicle's position, velocity, and acceleration as functions of time are defined. Reference frames are also needed to describe the vehicle's attitude relative, for instance, to points on the earth, to the sun, or to guide stars. Similarly, the time derivatives of the vehicle's attitude, particularly angular rates and angular accelerations, are often needed. Implied in the use of reference frames is also the existence of a precisely defined, "uniformly" evolving time. Easily defined reference frames are those that are rigidly attached to the earth, to some other astronomical body, or to the s t r u c t u r e - - a s s u m e d r i g i d - - o f the space vehicle itself. This latter frame is useful, for instance, in studying the dynamic effects on the vehicle resulting from the deploym e n t of solar panels, antennas, or scientific probes. Although easily defined, these reference frames will generally be accelerating; that is, they may be rotating or may also accelerate by parallel translation. Table 1.1 contains approximate values of the accelerations of several of such reference frames, expressed by their ratios to the standard gravity go = 9.80665 m/s 2. The first few accelerations listed in Table 1.1 are large and must be taken into account, together with other residual accelerations of comparable magnitude, in the planning of typical space missions. Particularly sensitive to small residual accelerations are missions in which a satellite is required to
2
C H A P T E P, 1 Reference Frames and Time
Table 1.1 Magnitudes of Various Residual Accelerations Acceleration at the equator due to the planet's rotation Earth Mars Acceleration on the earth's surface Due to earth-moon mutual attraction, with moon at zenith Due to sun-earth mutual attraction, with sun at zenith Acceleration (at the solar system location) due to rotation about the galactic center
3.44 10 -3 go 1.74 10 -3 go 1.13 10 -7 go 0.53 10 -7 go 2.24 10 -11 go
execute, during its lifetime, a large n u m b e r of orbits about a planet or moon. Unless corrected, small errors will tend to accumulate and will lead to large deviations from the i n t e n d e d path. On the other hand, the effect produced by the rotation of the solar system about the galactic center as a part of a spiral a r m of our galaxy is sufficiently small to be negligible for present purposes. Newton's second law refers only to inertial reference frames, that is, nonaccelerated frames. This t h e n raises two questions: How can an inertial frame be found? And how is one to relate m o t i o n s that are best described relative to an accelerating f r a m e - - s u c h as a frame attached to a spinning s p a c e c r a f t - - t o the same m o t i o n relative to an inertial frame? An i m p o r t a n t postulate of Newtonian mechanics asserts that once an inertial frame has been established, all other frames that move relative to this frame with constant velocity and no rotation are also inertial frames ("Galilean invariance") [1, 2]. Attempts to define inertial frames and time as those for which the laws of Newtonian m e c h a n i c s are valid are in essence circulatory. Definitions that attach inertial frames to the "fixed stars," taking advantage of the near-zero a p p a r e n t rotation of distant stars, leave o p e n the question of acceleration by parallel displacement. The m e a n i n g of an inertial reference frame (often referred to as "inertial space") in Newtonian mechanics can be m a d e clear as a limiting case in the theory of general relativity. Newtonian m e c h a n i c s is a model of space and time that postulates a h o m o g e n e o u s , isotropic space and a uniform time proceeding at a uniform rate. It is a sufficiently accurate representation for nearly all p u r p o s e s that are of interest to space technology engineers. Confining oneself to classical, Newtonian mechanics, one has at least the heuristic fact that no contradiction has ever been found by postulating the existence of an inertial space. In particular, no such contradictions have been found by observations of spacecraft paths. The a s s u m p t i o n that a frame of reference that is n o n r o t a t i n g with respect t o t h e m o s t distant stars, and is not accelerated with respect to the center of our galaxy, is for all p u r p o s e s a m o r e t h a n sufficient a p p r o x i m a t i o n to an ideal inertial space in Newtonian mechanics. Similarly, no contradiction has been found by ass u m i n g that the time in Newtonian m e c h a n i c s is the same as that m e a s u r e d by the m o s t reproducible timepieces, particularly by atomic clocks.
1.1 ReferenceFrames
3
Figure 1.1 Ecliptic and celestial equatorial planes. 1.1.1
The Ecliptic and Celestial Equatorial Planes The ecliptic (the path of the earth about the sun) is said to be in the ecliptic plane, which therefore contains the sun's and the earth's mass centers (Fig. 1.1). One also speaks of the celestial equatorial plane, which is the plane parallel to the earth's equatorial plane and through the sun's mass center. The ecliptic and celestial equatorial planes intersect, as illustrated in the figure, in a line referred to as the equinox line (because, when the earth on its annual path crosses this line, day and night have equal length). The crossing points are called the vernal equinox (the earth is at this point on about the 21st of March) and the autumnal equinox (about the 22nd of September). As discussed in Chapter 7, the vernal and autumnal equinox points play an important role in the operation of geosynchronous and other spacecraft. During two periods each year, centered around these points, and centered around local midnight, the sun will be eclipsed for these satellites. Special operational procedures are then needed to compensate for the lack of solar radiation. The direction of the equinox line that points from the earth at the vernal equinox toward the sun is referred to, for historical reasons, as the first point ofAries and is often designated by the symbol ~ . (As the name suggests, the line was found by early astronomers some 2500 years ago to point toward the constellation Aries. But since it moves by about 0.8' per year, it is presently in Pisces, moving into Aquarius.)
C H A P T E R 1 ReferenceFrames and Time
4
The ecliptic plane is inclined to the equatorial plane at an angle ic called the obliquity of the ecliptic. At present it is 23.44 ~ (Primarily as a consequence ofthe gravitational attraction bythe sun and the m o o n on the earth's equatorial bulge, the earth's spin axis, and hence also the equatorial plane, precesses relative to the most distant stars with a period of 25,920 years. The obliquity fluctuates between 21.5 ~ and 24.5 ~ with a period of about 41,000 years and is presently decreasing by about 0.5" per year. The eccentricity of the earth's orbit about the sun also has a small fluctuation, with a period of about 100,000 years. All these effects, however, are too small to be significant for most space operations.) 1.1.2
Reference Frames and Coordinate Axes
The following reference flames and coordinates are in frequent use [3]. They are arranged in order of increasing closeness to ideal inertial frames. (la) Earth-fixed geocentric equatorial frame: As shown in Fig. 1.2, this is the frame of reference conventionally used in cartography. Longitude is measured eastward from the Greenwich meridian (the former site of the Royal Observatory in Greenwich, a suburb of London). The latitude is measured positive northward, negative southward, starting at the equator. These two spherical coordinates determine the location, for instance, of a ground station. They are also useful in stating the instantaneous location of a near-earth space vehicle, by radially projecting its location in space on the earth's surface (a Azimuth
Xg3 Polar AxisB Elevation
Colatitude
/
Greenwich Meridian
~
.. "
Xg1"
I
f
/Xt3
i
Xtl
f
V ~ ~
~
/--/----- Equator
Xg2
/
Longitude
j
Unit Sphere
Latitude
Figure 1.2 Earth-fixed geocentric equatorial (Xgl, Xgz, Xg3)and topocentric (Xtl, xt2, xt3)reference frames. P, space vehicle location; O, ground station.
1.1 ReferenceFrames
5
spherical approximation to the geoid is assumed here). The curve obtained by this projection of a space vehicle moving relative to the earth is often referred to as the vehicle's ground track. The same system of spherical coordinates is also useful for other planets and for moons. The reference frames are then referred to as "planetocentric" and "selenocentric," respectively. Because longitude and latitude are angles i n d e p e n d e n t of the radius of the astronomical body, it is convenient to think of t h e m as arc lengths on the "unit sphere" (the sphere of radius one). (lb) Topocentricframe: The line of sight to the instantaneous position of a space vehicle is given by its azimuth and by its elevation (Fig. 1.2). The azimuth angle is measured in the local horizon plane ~ that is, the plane tangent to the (spherical) earth at the observer's l o c a t i o n ~ and starts at local north, positive for the direction toward local east. The elevation angle is zero at the horizon and positive for lines of sights above it. With the possible exceptions of short-range sounding rockets or tactical missiles, the two reference systems (la) and (lb) listed here are hardly ever suitable in space technology as approximations to an inertial frame. They are needed, however, for specifying space vehicle locations in coordinates that are directly observable. Coordinate transformations, to be discussed later, will therefore often be required to convert from one system to another. (2) Geocentric equatorial reference frame: In contrast to frame (la), this frame does not co-rotate with the earth. The origin of the spherical coordinate system is in the earth's mass center (or, correspondingly, in the mass center of other planets or moons). As shown in Fig. 1.3, the longitude is again Meridian
x~.~
/
/ P
PolarAxis
,'X /
, - eclinat,on
I / I / I / I / I/
..J.
~--/- Equator EquinoxLine Right Ascension"
~
~
~
Xg, 2
Unit Sphere
Figure 1.3 Geocentric, equatorial, nonrotating reference frame (Xg,~,Xg,2,Xg,3). For many space missions this frame is a sufficient approximation to the ideal inertial reference frame. P, location of space vehicle.
6
C H A P T E R 1 Reference Frames and Time Celestial North Pole Ecliptic North Pole
Xc3 Xc2 Ecliptic Latitude
\ \ \
q2 Ecliptic Longitude
Celestial Equatorial Plane Ecliptic Plane Right Ascension
Declination Unit Sphere
Xcl
Equinox Line qP
Figure 1.4 Heliocentric equatorial (Xq~,Xq2,Xq3) and ecliptic (Xc~,Xc2,Xca)reference flames. P, location of space vehicle.
taken positive eastward along the equator, but starts at the first point of Aries (equinox line). It is referred to as the right ascension of a point. The latitude is now referred to as the declination. This reference frame, which is very nearly nonrotating, is a suitable approximation to an inertial frame for the analysis of near-earth space vehicle trajectories and for their attitude control. (3) Heliocentric equatorial referenceframe: This reference flame, illustrated in Fig. 1.4, is the same as the one in item (2), except that the origin of the coordinate system is now in the mass center of the sun and the celestial equatorial plane now replaces the usual equatorial plane. This flame is an even closer approximation to an inertial frame than is the geocentric equatorial reference flame. (4) Heliocentric ecliptic referenceframe: Again, as is also illustrated in Fig. 1.4, the origin of the spherical coordinate system is in the sun's mass center. The ecliptic plane now replaces the celestial equatorial plane. Longitude and latitude are now referred to as the ecliptic longitude and ecliptic latitude, respectively. Again, the ecliptic longitude is measured eastward from the first point of Aries. The ecliptic latitude is taken positive extending toward the ecliptic north pole, negative when pointing toward the ecliptic south pole. This reference flame is most often used as a suitable approximation to an ideal inertial flame when analyzing space missions to the planets. (5) Barycentric referenceframe: This frame, which is a still closer approximation to an ideal inertial frame, has its coordinate origin in the center of mass of the solar system. It accounts, therefore, for the acceleration of the sun by
1.2 Motion in Accelerated Reference Frames
7
the planets, principally by Jupiter, and to a lesser extent by Saturn. Because of the large mass of the sun relative to the planets (Sun/Jupiter mass ratio approximately 1047), the barycenter stays at all times close to the sun, at a distance from the sun's center that is comparable to the solar radius. A nearly perfect inertial frame can then be constructed by using in place of the ecliptic plane the Laplaceinuariable plane, which is the plane through the barycenter and perpendicular to the orbital angular m o m e n t u m of all the masses in the solar system. The inclination of this plane is intermediate between the orbital planes of Jupiter and Saturn. Its inclination relative to the ecliptic plane is 1.65 ~. Although it is important in astronomy when calculating planetary positions over long time spans, either forward or backward in time, for the m u c h shorter times of flights of spacecraft within the solar system, the slight deviation of the nonrotating heliocentric ecliptic reference frame from an ideal inertial frame can almost always be neglected.
1.2
Motion in Accelerated Reference Frames The need to consider two or more reference flames that are in relative motion to each other arises frequently in all branches of dynamics. It is also important in space technology. An example is the relation between an inertial flame and a reference frame attached to a spin-stabilized spacecraft. A second example, in which both frames are accelerated, is the relation between an earth-fixed reference flame and the spacecraft-fixed flame. In analyzing, for instance, the effect of the sloshing motion of propellant in a vehicle tank, it will be advantageous to formulate the equations of motion of the fluid and the boundary conditions in a reference flame that is rigidly attached to the propellant tank, hence to the vehicle, yet taking into account the space vehicle's own translation and rotation. We consider here velocities and accelerations, hence derivatives of the position vector of a point with respect to time. Of interest here are reference spaces that are in motion relative to each other. As illustrated in Fig. 1.5, the reference spaces are indicated by their Cartesian coordinate axes. The coordinate flame (Xl, x2, x3) translates and rotates relative to the (X1, X2, X3) flame. The first flame's motion relative to the second one is determined by the time derivative of the position vector R0(t) and by the relative angular velocity vector ,~(t). The position vector r(t) defines the point P in the (Xl, x2, x3) coordinate flame. In applications to space technology, this point is often the instantaneous center of mass of a space vehicle. (The fact that the mass itself may be a function of time as a consequence of expenditure by the vehicle of propellant is immaterial for what follows.) Time derivatives in the space spanned by X1, X2, X3 will be designated by d/dt, in the space spanned by xl, x2, x3 by the dot notation ('). Let ui, i = 1, 2, 3 designate the orthonormal base vectors of the (x~, x2, x3) coordinate flame and b(t) an arbitrary vector with components bi(t) in this flame. Hence
b = ~ biui i
8
C H A P T E R 1 Reference Frames and Time
P
b
X~ R
r
/
dt
Figure 1.5 Illustrates derivation of Eqs. (1.1) to (1.3). By definition
I~l i - - 0 ,
whereas dui / d t = w x ui
so that
i, = Z
b,,,,
i
=
E ab, i
(bi being a scalar, there is of course no distinction b e t w e e n the two derivatives). From db/dt = ~[(dbi/dt)ui i
+ bi(w x ui)]
then follows db/dt-
b +w x b
(1.1)
This result is basic for relating the two time derivatives d b / d t and b of an arbitrary vector b. The transformation equation for the velocity of a moving point with position vectors r(t) in the (xl, x2, x3) coordinate flame a n d R(t) in the (X1, X2, X3) coordinate flame is obtained by taking r for b. Hence, with R = Ro + r, dR/dt = dRo/dt +/, + w x r
(1.2)
Similarly, the t r a n s f o r m a t i o n for the acceleration of the moving point is obtained by differentiation of (1.1) and identifying b with r and then with w x r. Thus the following result for transforming accelerations between two different reference spaces is obtained: d 2 R / d t 2 = d2Ro/dt 2 + ~ + w x (w x r) + 2w x/" + d~ x r
(1.3)
In the last t e r m it is immaterial w h e t h e r the time derivative of the relative
1.3 Example: The Yo-Yo Despin Mechanism
9
angular velocity is taken in one or the other of the two spaces, since by (1.1), identifying b with w, d w / d t = (o. Equations (1.1) to (1.3) apply to any two reference spaces, because these relations are purely kinematic. If, however, as happens frequently in applications, the (X1, X2, X3) coordinate frame defines an i n e r t i a l space, where therefore Newton's second law in its familiar form (where the acceleration is taken relative to inertial space) holds, it then follows from (1.3) that m[: = F - m [ d 2 R o / d t 2 + .;
x
(w
x
r) + 2w
x i"
+ d;
x
r]
(1.4)
where m is the mass concentrated at the point P and F the force acting on it. The second term on the right comes from the acceleration of the origin of the (Xl, x2, x3) coordinate flame, the third term is the centrifugal force, and the fourth is the Coriolis force. (No particular n a m e has been given to the last term.) The last four terms jointly are referred to as i n e r t i a l forces. They must be added to the force F to complete all forces acting on m in a noninertial space. In the next section, we discuss an application of (1.4) to space technology.
1.3
Example" The Yo-Yo Despin Mechanism Both in orbit and during orbit insertion, spacecraft and upper stage vehicles are frequently provided with spin to ensure attitude stability. At other times it may be necessary to terminate the spin permanently or at least temporarily, for example, to reorient the vehicle prior to a propellant burn. Despinning can be accomplished by firing small retro-rockets m o u n t e d on the vehicle such as to produce a torque counter to the direction of spin. Alternatively, despin can be obtained by initiating the so-called yo-yo m e c h a n i s m discussed in the following. This type of despin is frequently used in the case of cylindrically shaped vehicles. It has the advantage of great simplicity and reliability. As indicated in Fig. 1.6, one or several weights are attached to cables that are initially wrapped around the vehicle, typically in the plane through the vehicle's center of mass and normal to the spin axis (which is also one of the principal axes of inertia). Initially, the weights are attached to the body of the vehicle. Despin is initiated by releasing the weights pyrotechnically. As the cables now unwrap, the weights swing out on a path that is an involute of a circle when viewed in a reference flame corotating with the vehicle. The sense of the initial wraparound is such that the pull of the cables reduces the vehicle's angular m o m e n t u m . When the cables are fully extended, a split hinge releases them and the weights. The following analysis applies to a system with two symmetrically arranged weights, each of mass m. (X1, X2) designates an inertial reference flame. The flame (Xl, x2) with the corresponding unit base vectors Ul and u2 is moving with the instantaneous location of the tangent point O of the cable. The angular velocity of the vehicle is designated by fl(t), and ~z(t) is the angle between the weight attachment/release point and the cable's
10
C H A P T E R 1 Reference Frames and Time
m
Cab
x2
~---
Path of m relative to spacecraft
Ul
u2
" - - - - - - Attachment/release of m X1 Attachment/releaseof cable Spacecraft
Figure 1.6
Schematic of despin yo-yo mechanism.
tangent point. Hence the angular velocity of the (X1, X2) reference frame is
~=~+~
(1.5)
The free length of the cable is Ro ~, where Ro is the radius of the spacecraft. The position vector for the point mass m in the (x~, x2) flame is r Ro~kU~, a n d R - Ro(~ku~ + u2) in the (X~, X2) system. From (1.2) then follows dR dt
=
d~
Ro ---d~Ul-t
dUl
du2 )
' ~ , r - - ~ -{-- ---~-
--
R0(6Ul -l- ~WU2 -- O)Ul) = R0(-~Ul + 'I/zogU2)
(1.6)
The total a n g u l a r m o m e n t u m L of the vehicle and the two masses is in the direction of the X3 axis and is given by L = I3f~ + 2 m ( R x d R / d t )
(1.7)
where/3 is the vehicle's m o m e n t of inertia about the spin axis. Hence
L = I3fl + 2mR2(~2o9 + ~) With the initial condition ~ = 0 at t = 0, it follows from the constancy of L that k(~O-
~ ) "-
~r20)
(1.8)
1.3 Example: The Yo-Yo Despin Mechanism
11
where
f2o = f Z ( t - 0),
k - I3/(2mR 2) + 1
(1.9)
The total kinetic energy T of the system is 1 ~'22 + 2 m ( d R / d t ) 2) __ 213 1 ~,'22 + T = 2(13
mR~(g22 Jr- ~r2092)
(1.10)
From the constancy of T follows k(C~ - ~2) = V,2~ ~
(1.11)
and d i v i d i n g by (1.8), ~o + f2 = w = f~ + V}. Therefore ~ = ~o = const, so that ~k = f2ot. F r o m (1.8) and (1.5), k(~o - f2) = ~ 2 ( ~ o + ~ ) t a. Solving for then gives f2 b e tw e e n t = 0 and the final t i m e tf = //(Rof2o) where / is the length of each cable. The final result, therefore, for the vehicle angular velocity at some t i m e t is
k - ~2~tz k + ~22t2'
S2 - f~o ~
l - Ro~20
0 < t <
-
(1.12)
Solving instead for l, one obtains an expression for the line length required to despin the vehicle from an initial angular velocity ~20 to the final velocity ~f:
l--Ro(k~~
~20 + ~ff
If a complete despin to s l--
1/2
(1.13)
= 0 is required, t h e n I = Rok ~/2, that is,
(,3
~-~+
,
~'2f -- 0
(1.14)
This result is remarkable because it shows that the cable length for complete despin is independent of the initial angular velocity of the vehicle. Deviations from the assumed initial velocity therefore do not propagate as errors affecting the intended final zero rate of spin. As an example, for a fairly typical upper stage vehicle, with/3 - 1000 kg m 2, R0 = 1.00 m, and m = 1.00 kg, the required length of each of two cables is 22.4 m, corresponding to about 3.5 turns around the vehicle drum. The preceding expressions used for the angular m o m e n t u m and kinetic energy assume that the vehicle is a single rigid mass. Liquid propellants in the spacecraft, however, have their own dynamics, and their rate ofrotation during the spin-down may be different from that of the vehicle. Nevertheless, these effects can be minimized by subdividing the propellant stores into several, eccentrically arranged tanks. When a new vehicle is being developed, it is customary to check the analytical results by an actual spin-down test on a spin table. A disadvantage of the yo-yo system is the fact that the free-flying weights and cables could potentially cause a collision with other spacecraft in a similar orbit. This possibility cannot be entirely dismissed
12
C H A PT E R
1 ReferenceFrames and Time
in the case of orbits that are heavily populated, such as is the case for the geostationary orbit.
1.4
Euler Angles and Transformations of Coordinates To describe the orientation of a three-dimensional Cartesian coordinate system with respect to another such system, in general three angles are needed. These then can serve as generalized coordinates. Different choices are possible, but choices usually most convenient in applications to spinning rigid bodies, and therefore also to space vehicles, are the Euler angles r O, ~. To define them, we consider the Cartesian coordinate systems (X~, X2, X3) and (Xl, x2, x3) as illustrated in Fig. 1.7. Three rotations through angles r O, ~/, (in this order) will rotate the X1, X2, X3 axes first into the intermediate x~, x~, x~ axes, these then into the x~', x~', x~' axes and these finally into the Xl, x2, xa axes. (Caution is n e e d e d because in some texts on rigid body dynamics, the definitions of 0 a n d r are opposite to those used most often in space technology and also used here.) The 0 is referred to as the precession X3,x3"
x3",x3
x2
~1
X2" X2
x~ X1
Xl
\ Xl" ' Xl""
x~
X2"
X2"" X2
1"
P
X2"" P
X3"
x )(3""
[R~]
Figure 1.7
[R0]
[R]
Coordinate system rotation defined by Euler angles.
1.4 EulerAngles and Transformations of Coordinates
13
angle, 0 as the nutation angle, and ~ as the spin angle. The first rotation, Rr is through the angle ~, the second, R0, through the angle 0, the third, R~,, through r In matrix notation, these transformations about single axes are
Ixl [xa1 4
x3
4'
x;
X1
where [Ro]=
X1
[1 I I X2
where [Re] -
--[ R~/,]
x~'
where [Rv,]=
I cos~ -sin4~ 0 1 0 0 cos 0 0 -sin0
I
I cos~k -sine 0
0]
sine cos4~ 0 0 1
(1.15a)
0 ] sin 0 cos0
(1.15b)
sine cos~ 0
0] 0 1
(1.15c)
as is easily verified by the geometric relations read from the figure. The general transformation of the coordinates of a point P, or, what comes to the same, of the components of a vector, is obtained by the matrix product [R~,][Re ] [Re] so that
[xxj EXl] x2 x3
- [R]
X2 X3
where [R] = [R,][R0][R~]
(1.16)
Calculation of this product results in
[R]-
I cr162 sec0sr -cesr162162 s~sO
s~cr +c0sr -sesr162 -c4,sO
s0sr s0cr cO _1
(1.17)
where s stands for the sine and c for the cosine. Although complicated, this matrix plays an important part in much of the software needed for the attitude control of space vehicles. All matrices (with the exception of (1.19)) in this section belong to the class of orthogonal matrices. Mathematical properties of orthogonal matrices that play a role here may be summarized as follows: (1) Real valued, nonsingular matrices [A] are orthogonal if and only if [AT] [A] = [A] [AT] = [I] (where [ IT indicates the transpose and [I] the identity matrix). It follows from this that [A] -1 = [A] T. If [A] is orthogonal, so are [A]T and [A] -1. If [A] and [B] are orthogonal, the same is the case for the matrix product [A] [B]. (2) The determinant of an orthogonal matrix equals either +1 or -1. If +1, the orthogonal matrices effect the transformation of a right-handed (lefthanded) orthogonal coordinate system into another right-handed (lefthanded) orthogonal coordinate system. Products of orthogonal matrices
14
C H A P T E R 1 Reference Frames and Time with determinant + 1 are again of the same type. Examples are the rotation matrices [Re ], JR0], [R~ ], and [R] in this section.
[xlI Exll
The transformation inverse to (1.16) is therefore X2 x3
=
[ R T]
x2 x3
where [RT] = [R~][R~] [R~]
(1.18)
hence simply obtained by taking the transpose of [R]. Ifthe coordinate system (Xl, x2, x3) rotates with respect to the (X1, X2, X3) system with the angular velocity w (t), this can be expressed either by the temporal derivatives r ~, ff of the Euler angles or by the components of w in one or the other of the two coordinate systems. Most useful in applications to spacecraft is to express w in terms of its components in the (Xl, x2, x3) system, which here is assumed to be rigidly attached to the spacecraft. From the geometrical relations shown in Fig. 1.7 and the transformation equations, one finds for these components [Wl] [sin0sin~k we = s i n 0 c o s ~ w3 cos0
cos~k -sin~ 0
01 I r ] 0 ~ 1
(1.19)
The matrix associated with this transformation is generally not orthogonal. A difficulty arises in the use of Euler angles when 0 is approximately zero. In this case, the two planes shown in Fig. 1.7 nearly coincide, with the consequence that the nodal line (the x'~and x~' axis) is now poorly defined. In computations performed by a spacecraft computer, this will be evidenced by the subtraction of pairs of large, nearly equal numbers, leading to large errors. During normal operation of the spacecraft, it will usually be possible to choose the orientation of the axes such that 0 always remains fairly large. This, however, will not solve the problem that may arise, for instance, when a spacecraft may be tumbling out of control and must be brought back to a stable attitude. In this case, all possible values of the angles may occur. This problem of indeterminacy of the nodal line can be avoided by introducing more than three, hence interdependent angles, for example, by using two Euler angle systems with their x3 axes at 90 ~ to each other. Guidance computer software then can provide for switching from one to the other subprogram depending on the magnitude of 0.
1.5
Time Intervals and Epoch For precise definitions of time, one needs to distinguish between time intervals between two events and the time of a single event as conventionally found from a calendar and a clock. This latter time is technically referred to as the epoch of the event. A highly precise definition and standard for time intervals is needed, for instance, for radar measurements that support spacecraft navigation or altimetry. Thus, to determine the distance of a spacecraft from a ground
15
1.5 Time Intervals and Epoch
station by a one-way microwave signal p r o p a g a t i n g at the speed of light, the precision of the time interval m e a s u r e m e n t m u s t be of the order of 3 ns if a position accuracy of 1 m is to be achieved. Examples of the epoch of an event are the time at which a rendezvous of two space vehicles occurs a n d the time of a launch. Because, historically, the accuracy of time m e a s u r e m e n t s has greatly improved with time, ever more precise definitions of time have b e e n introduced. M a n y of the older definitions are still in use, however, in part because i m p o r t a n t astronomical data are stated in these terms [4]. An ideal system of t i m e k e e p i n g would be one in which time elapses perfectly uniformly, in the sense that the laws of physics are invariant to the time w h e n the physical event takes place. 1.5.1
International
Atomic
Time
(TAI)
Since 1967, the accepted s t a n d a r d unit for time intervals has b e e n the Syst~me I n t e r n a t i o n a l (SI) second. It is defined as equal to 9,192,631,770 periods of a certain atomic r e s o n a n c e frequency, in the microwave range, of cesium 133. A large n u m b e r of atomic standard clocks exist worldwide. To achieve great accuracy, m e a s u r e m e n t s from these clocks are c o m p a r e d and c o m b i n e d into an international standard. There also exist other types of atomic clocks, as indicated in Fig. 1.8 from Ref. 5. C o m p a r i s o n of such clocks shows that they tend to have slightly different drift rates, some short t e r m a n d s o m e longer term. For an averaging time of 1000 seconds the s t a n d a r d deviation of cesium or r u b i d i u m clocks
1E-9
i
i
i
i iiii
I
_ _ _
........
I
........
I
........
I
........
I
.....
~-
1E-10 _
1E-11 -c. mO
7 1E-12 1"13
~
High-performance ~"~~7~bidium 7
__
-~_~
High-performance -
"O "0
-o 1E-13 c-
t~
1E-14
~
a
s
e
Hydrogen 7
r
1E-15 1E-16
........
0.01
I
O. 1
........
I
1
........
I
10
........
Averagingtime (s)
I
100
........
I
1,000
......
10,000
Figure 1.8 Performance of atomic time 'standards. From Suder, J., Ref. 5, in Pisacane V. L. and Moore, R. C., eds., "Fundamentals of Space Systems." Copyright 9 1994 by Oxford University Press, Inc.. Used by permission of Oxford University Press.
16
C H A P T E R 1 Reference Frames and Time
is a few times 10 -13, corresponding to an error of about 1 second in 100,000 years [5]. Accuracy of this order is sufficient to observe, for instance, the slight irregularities in the rate of rotation of the earth such as caused by the tides, ocean currents, or variations of the mass of the polar ice caps. Hydrogen masers are capable of still higher precision but are relatively bulky. The epoch has been c h o s e n so that it is equal to UT1 on 0 hour on 1 January 1958. (UT1 is one of a family of the so-called Universal Time (UT) standards, discussed in the following.) Atomic clocks on satellites are being used for worldwide, highly precise position determination and navigation. Thus the United States' Global Positioning System (GPS), which is based on 24 satellites in three different orbits, uses atomic clocks in each of the spacecraft and also in the groundbased control and c o m m a n d center. Using portable receivers, positions on the ground can be determined to an accuracy of about 10 m. For purposes of surveying, centimeter accuracies of relative positions of two points can be obtained by taking advantage of a comparison of the carrier frequency phases received by the two points and by long integration times.
1.5.2
Universal Time (UT) and Greenwich Mean Solar Time (GMST) In astrodynamics, for both historical and practical reasons, Universal Time is most often used. Closely related to it by a mathematical formula is the Greenwich Mean Solar Time. These systems of timekeeping are derived from observations of the sun's crossing the Greenwich meridian at noon. A solar day is the time interval between two successive crossings by the sun of the observer's meridian. Primarily because of the eccentricity of the earth's orbit about the sun, the length of the solar day is not constant (the earth is closest to the sun in early January, most distant in early July). Solar days therefore must be averaged to serve as a basis for m o d e r n timekeeping. This averaging is accomplished by introducing a fictitious m e a n sun that moves uniformly along the celestial equator. Time measured in this way eliminates the irregularities caused by the eccentricity of the earth's orbit and by the lack of coincidence of the equatorial and ecliptic planes. It retains, however, the small irregularities caused by variations of the rate of rotation of the earth and of the location of the earth's poles. The resulting system of time is referred to as Universal Time. The term Greenwich Mean Solar Time is frequently used as synonymous with Universal Time, although, in principle, there is a small difference that manifests itself over centuries, a difference that is u n i m p o r t a n t in space technology, considering the much shorter time spans of space missions. Except for the addition or subtraction of an integer n u m b e r of hours, Standard Time or Zone Time, used in everyday life, is the same as GMST. There are several versions of the system referred to as Universal Time. To some extent, they depend on the earth's rate of rotation and on pole wandering. The latest and most precise standard is referred to as the Coord i n a t e d Universal Time (UTC). It is now based on TAI but differs from it by an integer n u m b e r of seconds, called leap seconds. Such leap seconds are introduced into UTC on 1 January or 1 July whenever needed to maintain
1.5 Time Intervals and Epoch
17
consistency within 0.9 second between the atomic time reckoning and a system that is based on astronomical measurements. 1.5.3
Sidereal Time Similarly to UT and GMST, sidereal time (Latin sidus: star) is based on astronomical determinations. It uses the vernal equinox as a reference point. A sidereal day is therefore the time interval between two successive passages of the vernal equinox across the observer's meridian. Sidereal time is internally consistent with the geocentric equatorial reference flame discussed in Sect. 1.1.2. Because of variations in the rate of rotation of the earth, sidereal time is not exactly uniform. However, for m u c h of space technology, its precision is entirely sufficient. Sidereal time and still other time systems referring to the solar system are often convenient for purposes of analysis. In applications to space technology, the results of such calculations are then usually translated into either TAI or GMST. Important in space technology is the difference in length between a sidereal day and a solar day, as illustrated in Fig. 1.9. The direction ofmotion of the earth on its orbit is the same as that of the earth's diurnal rotation. Both are counterclockwise w h e n viewed from the north. The earth moves each day through a heliocentric angle of 360 ~ divided by 365.2 m e a n solar days, which a m o u n t s to approximately 1 degree per day. As it takes the earth about 4 minutes to rotate through this angle, the solar dayis about 4 minutes longer than the sidereal day. A more accurate calculation that takes into account the obliquity of the ecliptic plane shows that the length of the sidereal day is 23 h 56 m 4.1 s.
Figure 1.9 Schematic illustrating the difference between sidereal and solar day.
18
C H A P T E R 1 Reference Frames a n d Time
Reference 5 contains a more thorough, yet concise discussion of the various systems of timekeeping. A more extensive treatment may be found in Ref. 3. Practical details may be found in the Explanatory Supplement to the Astronomical Almanac [4].
Nomenclature external force standard gravitational acceleration obliquity (inclination) of the ecliptic (Figs. 1.1, 1.4) m o m e n t of inertia about x3 axis constant [(Eq. 1.9)] length of cable angular m o m e n t u m mass position vectors (Fig. 1.5) time kinetic energy orthonormal base vectors (Fig. 1.5) Euler angles (in Sect. 1.3, ~ denotes the angle shown in Fig. 1.6) angular velocities
F
go - 9.80665 m / s 2 ic
13 k I L
m r, R t T
Ui ~,O, r ~,f2
Problems (1) At some specified time the position of a spacecraft is at a right ascension of 30.00 ~ and a declination of 60.00 ~ The obliquity (inclination) of the ecliptic is 23.45 ~. (a) Find the ecliptic longitude and latitude of the position by using spherical trigonometry. (b) Obtain the same result byusing the transformation ofcoordinate systems based on Euler angles. (2) Let ~(t) be the angular velocity ofthe orthonormal coordinate frame (Xl, x2, x3) relative to the orthonormal coordinate flame (X1, X2, X3). Let 091,092, 093 be the xl, x2, x3 c o m p o n e n t s of ~. (In m a n y applications to the control of the attitude of space vehicles, the first coordinate flame would be fixed to the vehicle, the second flame would be an inertial flame. However, the formulation of this problem is purely kinematic, hence independent of the assumption of an inertial flame.) (a) Defining the Euler angles as shown in Fig. 1.7, show that for 0 ~ 0 I~l 6~ 7}
-
I sin~/sin0 cos 7~ - s i n ~p/tan 0
COS~ / s i n 0 0 l
-sin ~ - c o s 7r/tan 0
0 1
I0911 coa O93
Note that this matrix is n o t orthogonal. (b) Show that this matrix is, as expected, the inverse of the matrix in Eq. (1.19).
References
19
References 1. 2. 3.
4. 5.
Goldstein, H., "Classical Mechanics," Addison-Wesley Publishing Company, Reading, MA, 1950. Whittaker, E. T.,'Analytical Mechanics," Cambridge University Press, London, 1937. Kovalevsky, ]., Mueller, I. I., Kolaczek, ]., eds., "Reference Frames in Astronomy and Geophysics," Astrophysics and Space Science Library, Vol. 154, Kluwer Academic Publishers, Dordrecht, 1989. Seidelmann, P. K., ed., "Explanatory Supplement to the Astronomical Almanac," University Science Books, Mill Valley, CA, 1992. Pisacane, V. L. and Moore, R. C., eds., "Fundamentals of Space Systems," Chapter 3 by Black, H. D. and Pisacane, V. L. Oxford University Press, New York, 1994.
a This Page Intentionally Left Blank
2 Forces and Moments The principal forces that act on space vehicles are gravity, thrust, and, within planetary atmospheres, also aerodynamic drag and lift. Thrust is derived from rocket motors and is an essential feature oflaunch vehicles and u p p e r stage vehicles. On a m u c h smaller scale, thrust is also used on artificial satellites and other spacecraft for position and attitude control. Aerodynamic forces are important for launch vehicles. They are also important for space vehicles that operate in a planetary atmosphere. This is particularly so in the case of reentry vehicles, that is, recoverable vehicles that reenter the earth's atmosphere on their return from a space mission. Aerodynamic f o r c e s ~ although often m i n u t e ~ m a y also result from the interaction of a spacecraft with the tenuous gas of the outer atmosphere. In the case of orbiting spacecraft, after repeated revolutions, the effects on position due to these forces tend to accumulate with time. They can then become highly significant in disturbing the orbit. The very small force that results from the solar radiation pressure can also play a role in spacecraft orbits. Solar sailing, which would use solar radiation pressure for thrust, is a distinct future possibility. Principally because of the difficult problem of deploying the very large sails required, solar sailing has not as yet been put into practice. Still other forces that are important in space vehicle engineering are mechanical reaction forces. An example is the force that results from the action of springs that are used to push apart vehicle stages after stage separation. Other examples are provided by the interaction of launch vehicles with their support during launch. For these forces Newton's third law holds. Associated with these forces are the corresponding m o m e n t s about some reference point. Most often, in the case of space vehicles, the center of mass is chosen as this reference point. As a consequence of the expenditure of propellant, the center of mass can shift relative to the vehicle's body. When this occurs, it is necessary to treat the vehicle as a system of variable mass and variable m o m e n t s of inertia. The theory of dynamics of a space vehicle is therefore generally more complicated than the classical theory of rigid bodies.
2.1
Gravity According to Newton's law of universal gravitation, the gravitational force Fg exerted by a point mass ml on a point mass m2 is Fg 21
G ml m2 = -
r------~-~r
(2.1)
22
C H A P T E R 2 Forces a n d M o m e n t s
where r is the radius vector that extends from the first to the second point mass. The constant G = 6.6732 10 -~1 m 3 /(kg s z) is known as the u n i v e r s a l gravitational constant.
If instead of a point mass, m l is the mass of an astronomical body, not necessarily spherically symmetric, then (2.1) is still valid asymptotically at distances large c o m p a r e d with the dimensions of the body. As discussed in Section 2.1.1, it also remains valid even at arbitrary distances, if the body's m a s s distribution is spherically symmetric. It is preferable to introduce in place of G and of m l their product, 13,--Gml, the so-called gravitation p a r a m e t e r of the astronomical body. The reason for this preference is that/z can be determined by astronomical and satellite m e a s u r e m e n t s m u c h more precisely than can G or m~ separately. In applications to astrodynamics, it is only their product that matters. Some examples of the numerical values of the gravitation parameter are: Sun: Earth: Moon:
~ - 1.32712 1011 klTl3/S 2 = 3.986006 105 km3/s 2 (GEM-L2 geopotential model, 1983) = 4.90265 103 km 3 / s 2
If one takes for m2 a unit mass, then (2.1) defines a force field--strictly speaking an acceleration, say, g(r) - - at points given by the position vector r. As is seen from the form of (2.1), this field is conservative and can therefore be expressed by the gradient of a potential, say, ~g(r), so that g = - g r a d ~g
(2.2)
(This definition of the potential with a negative sign is conventional in physics and in space technology, but usually not in astronomy.) It follows from (2.1) that if m l is a point mass, Og =
/z r
(2.3a)
If m~ is an extended mass, such as that of an astronomical body, one obtains by superposition of the effects of the elements of ml,
*g=
ml
fm dml
, I r - r'l
where r and r' are as defined in Fig. 2.1.
Figure 2.1 Illustration of Eq. (2.3b). O, center of mass.
(2.3b)
2.1 Gravity
23
9 g(r) is seen to have the dimension of an energy per unit mass and is equal to the gravitational potential energy per unit mass at r. The arbitrary constant that can be added to ~g conventionally is chosen such that the potential energy is zero at an infinite distance from ml. Hence it is negative at other distances. Deviations from an exact spherical mass distribution of an astronomical body result in a gravitational potential that is no longer perfectly symmetric. This can result in important long-term effects on orbiting spacecraft, particularly those that orbit at low altitudes. Conversely, observations of the orbits of near-earth artificial satellites provide the most precise method for determining the exact figure and mass distribution of the earth, of other planets, and of the moon. 2.1.1
Spherically Symmetric Mass Distribution If the density of an attracting body is such that it depends only on the radius, the expression for the gravitational potential is particularly simple. The result is applicable as an approximation to the earth, for instance, because the density, although much higher in the nickel-iron core than in the mantle, depends only weakly on latitude and longitude. The result, which was already known to Newton, can be obtained by integration over the elements of mass composing the body. The same result can also be obtained by the use of Gauss' integral theorem, a theorem that proves useful in other problems related to gravitational fields. For this reason, we use it here. Let S be a spherical surface of radius r, concentric with and outside the body. Also let n be the inward unit normal to S and dg the gravitational acceleration, caused by an element d m of the body, at the element d S of S. The distance between d S and d m is designated by r'. Hence from (2.1), n . d g = G d m c o s O / r '2
where 0 is the angle between the inward normal and the line from d S to d m Integration over S results in
fs
n . dg dS = G d m
/s c~
r' 2 d S - G d m
dS2 - 4rcG d m
where d~2 is the solid angle subtended from d m to dS. The result is seen to be independent of the location of d m within the body. Summing over all elements dm, we obtain the acceleration g at a point on S. By symmetry, g is normal to S and has constant magnitude on S. Therefore rag G m l = / z so that if r0 is the radius of the body g-
- t z r / r 3,
r > ro
(2.4)
a result that is identical with what follows from (2.1) for a point mass. To summarize: for a spherically symmetric body, its gravitational field outside it is the same as if the total mass were concentrated at the center. This field, which falls off with distance as the inverse square, is therefore an example of a central, i n v e r s e - s q u a r e f i e l d .
24
C H A P T E R 2 Forcesand Moments
2.1.2
Gravity Gradient Effect If, over the space occupied by a space vehicle, the external gravitational field were perfectly uniform, the resultant of the forces on the various parts of the vehicle would be simply the force obtained by placing the entire mass at the vehicle's center of mass. In most applications, the assumption that the gravitational field is uniform over the extent of the space vehicle is satisfied to very high precision. Exceptions, referred to as gravity g r a d i e n t effects, occur only w h e n the dimensions of the vehicle are relatively large or w h e n the time over which dynamical effects resulting from the nonuniformity can accumulate is very long. One can distinguish two such effects: (1) The resultant of the gravity forces that act on different parts of the space vehicle is not exactly the force obtained from the gravity at the location ofthe center ofmass. (2) The gravity forces produce a torque about the center of mass. Applications to space technology in which the first effect is important occur in the maneuvering of two space vehicles before docking, but hardly ever in the dynamics of a single vehicle. An exception may occur when an instrumented capsule is connected to the spacecraft by a long tether, an arrangement that is useful for in situ studies of the u p p e r atmosphere. The second effect, although m u c h too small to be significant for launch vehicles, upperstage vehicles, or most spacecraft, can be important as a perturbation of the attitude (spatial orientation) of very large space structures. Another application occurs in the gravity g r a d i e n t stabilization of near-earth satellites by means of a long b e a m that is attached to the spacecraft. In what follows, we consider as an example a near-earth large structure, such as a truss (Fig. 2.2), which may serve as a c o m p o n e n t of a space station. The gravitational field is assumed to be a central, inverse-square field.
r
P
/
/
I
I
/
I I
I
I
Figure 2.2 Gravity gradient effect. Illustration for Eqs. (2.5) to (2.9).
2.1
Gravity
2.5
We designate by R0 the radius vector from the center of the earth to the center of mass O of the structure and by R the corresponding vector to an element of mass dmat P. The m o m e n t about O induced by the gravitational forces is /,
Mg
]m r x (R/R 3) dm
(2.5)
where/z is the earth's gravitational parameter, m the mass of the structure, and r = R - R0. Since r << R0, it will be sufficient to consider the terms of lowest order in powers of r/Ro. From the figure, R 2 = R02 + r 2 + 2R0. r. Using the binomial theorem to expand the cubic term, R 3, in (2.5), one finds to the lowest significant order
Mg - - ~ fm r =
x (no + r)Ro3(l + 2110. r/R 2 + r2/R~) -3/2
R~ R0 x
fm
(R0. r)r
dm
dm
(2.6)
The term linear in r vanishes in the integration, since, by the definition of
the center of mass, fm r dm = O.
The integral can be expressed in terms of the m o m e n t s of inertia of the structure. For this purpose it is convenient to introduce a Cartesian coordinate system with axes that coincide with the principal axes of inertia. The c o m p o n e n t s of r and R0 will be designated by (X1, X2, X3) and (X01, Xo2, X03) respectively. From the definition ofthe m o m e n t ofinertia I1, w h e n expressed by the principal axes c o m p o n e n t s of the position vector r,
Ii-fm(X?a+x?3)dm,
fmXlX2dm=O, fmXxX3dm:O
Corresponding expressions for 12 and 13 are obtained by cyclic interchange of the indices of the components. By multiplying out the scalar product in (2.6), one obtains for the first c o m p o n e n t of (R0 9r)r (X01Xl -[- X02X2 "4- X03x3)Xl
and correspondingly, by cyclic interchange, for the other components. Therefore
(fm (Ro " r)r dm) l =X01
fmx21dm =
1X01 (12 -~- 13 -- 11)
and correspondingly for the other components. This result can be s u m m a r i z e d by introducing the Cartesian tensor l represented by the matrix
89 + I3- h)
o
o
0
1 (I3 --[--h -- I2)
0
0
0
l(h +/2-/3)
I
1 (2.7)
which, when postmultiplied with the vector Ro = (Xol, 202, X03) T, shows
26
C H A PT E R 2 Forces a n d M o m e n t s that
fm(
RO . r) r d m = J. Ro
(2.8)
The final expression obtained for the gravity g r a d i e n t m o m e n t , to the lowest significant order, is therefore 3/z Mg = - ~ R o x (]. Ro) n6
(2.9)
The following example is of interest because it illustrates the order of magnitude of the effect. To simplify the problem, we approximate the truss in Fig. 2.2 by a thin rod of length 21 and uniform mass O per unit length. The m o m e n t of inertia, transverse to the rod and about the center of mass, is then I = (2/3)013 = (1/3)m/2 where m is the mass. Let 0 designate the angle between the rod and the local vertical. Therefore Xol = Ro s i n O,
Xo2 - - O,
Xo3 ---- -- No c o s 0
Evaluation of the expression R0 x (J. R0) in (2.9) then gives the magnitude Mg of the gravitational torque as
/~rrd2 Mg -- 2R 3 sin(20)
(2.10)
The torque is seen to have a m a x i m u m at0 = 45~ For/z = 3.986105 km3/ s 2 (earth), R0 = 10,000 km (i.e., about 1.57 times the earth's radius), m = 100 kg, l = 10 m, the m a x i m u m torque affecting the rod becomes 1.99 1 0 - 3 Nm.
2.1.3
The Earth's Gravitational Field To predict the paths of near-earth orbiting spacecraft, a more exact knowledge of the gravitational acceleration than that afforded by (2.4) is needed. This can be obtained by a combination of geodesic and gravimetric measurements on the surface of the earth, which can then be used to infer the gravitational field evel~vhere externally to the earth. This is made possible by the application of a well-known theorem in potential theory, to be discussed later in this section. With the advent of space technology, observations of the paths of nearearth satellites have been complementing the e a r t h - b o u n d measurements. Similarly, spacecraft are being used to determine the gravitational fields of other planets or of the moon. The earth's gravitational field lacks exact spherical symmetry for two reasons: (1) the surface is not exactly spherical; (2) the mass density in the interior is not exactly symmetrically distributed. The presence of these asymmetries has important effects on the orbits of near-earth satellites. The largest contribution to the asymmetry is caused by the e q u a t o r i a l bulge, which is the result of the centrifugal force produced by the earth's rate
2.1 Gravity
27
of spin. Thus, whereas the polar radius (the radius taken along the earth's axis) is approximately 6357 km, the mean equatorial radius is 6378 km, larger by about 21 km. A useful concept, first introduced by Gauss, for describing the figure of the earth is the geoid. This is defined as the surface that would result if the earth were entirely covered by an ocean. Adding to the gravitational force the (much smaller) centrifugal force caused by the earth's rotation, the combined force for reasons of equilibrium must be perpendicular to the surface. The geoid is therefore an equipotential surface. An approximation to the geoid is the reference ellipsoid [1], which is usually taken to be the rotationally symmetric ellipsoid that best approximates the geoid, with its axis of symmetry along the polar axis. This reference ellipsoid, however, does not take into account, for instance, the variations of the force field with geographic longitude, nor such anomalous effects as the large gravitational perturbations over oceanic islands. (This latter effect, however, is not significant for calculations of the paths of satellites, because, being highly localized, it falls off rapidly with altitude and produces only a short-term interaction on passing spacecraft.) More important for predicting the paths of near-earth satellites and their slow drifts out of the original orbital plane is the gravitational influence ofthe moon and the sun. This topic is beyond the scope of the present section. The determination of the path of near-earth satellites, given the perturbations of the gravitational fields of the earth, moon, and sun, is treated extensively in Ref. 2. It is not considered in this book. A representation of the earth's gravitational field, more exact than is provided bythe reference ellipsoid, will occupy the remainder ofthis section. It follows from the form of Newton's universal law of gravitation that the divergence of the force field of a point mass is zero. By superposition of the effects of point masses, this result can be extended to the field exterior to a distributed mass, as is the case for the gravitational field exterior to the earth. Its potential (I)g therefore satisfies Laplace's equation (2.11)
V2 (I)g --0
We note first some mathematical preliminaries [3, 4]: If V2r = 0 and r is prescribed on a dosed surface S, then r is deter-
mined everywhere outside S (the Dirichlet problem). Similarly, if the component normal to S of grad r is prescribed on S, then r is determined everywhere outside S (the Neuman problem). In terms of spherical coordinates, Laplace's equation takes the form 1 8 (028~)
0200
-~0
1
8 (sinO8r
+ 0 2 sinOOO
O0
i
82r
(2.12)
+ 0 2 s i n 2 0 0)~2 --
where 0 is the distance from the origin, 0 the colatitude (in applications to geodesy, the angle extending south from the north pole, as indicated in Fig. 1.2) and the latitude )~. In these coordinates, Laplace's equation is separable. If one lets
r
=
R(O)|
(2.13)
28
C H A P T E R 2 Forces a n d M o m e n t s
t h e n the following three equations are obtained (m, n = constants)" 2 + m2A - 0
d2A/dk
1 d ( d O ) sinOdO s i n O ~
(
+
1 d (dR) 02
02 d o
-~0
n(n+l)
(2.14a) m2 ) sin 20 0 - 0
n(n+l) - ~ R= 0 02
(2.14b)
(2 14c)
The f u n d a m e n t a l solutions of (2.14a) are cos(mk) and sin(m~). The req u i r e m e n t of periodicity and continuity of A requires that m be an integer. Without loss of generality, we can take m > 0. Imposing the condition that O be finite in the range 0 < 0 < Jr, the f u n d a m e n t a l solutions of (2.14b) are the Legendre functions of the first kind Pnm(cos 0) with integer n a n d 0 < m < n. If m = 0, the Legendre functions are polynomials, usually written as Pn, omitting the superscript. The Legendre polynomials can be defined by successive differentiation as follows, with x - cos 0 ~ 1 dn Pn(x) -- 2nn---~.dx-----g(x 2 - 1) n
(2.15a)
From this the Legendre functions follow by Pro(x) -- (1 -
X2) m/2 d m p n ( x ) dx m
,
0 <_ m <_ n
(2.15b)
The f u n d a m e n t a l solutions of (2.14c) that fall off to zero as the distance from the coordinate origin b e c o m e s infinite are 1/0 n+ 1 (end o f m a t h e m a t i c a l preliminaries). One needs to add to the gravitational potential ~g of g the potential (I)f from the inertial forces that arise from the fact that earth-fixed coordinates are used rather than inertial coordinates. At the equator, the centripetal acceleration is approximately 0.0339 m / s 2, hence 0.00344 times normal gravity. At general points in near-earth space, as is easily seen, this additional potential is ~,'~2~O2 s i n 2 0 (I)f ---
-
-
2
(2.16)
with ~2 - 7.292115 10 -5 rad / s the rate of rotation of the earth. The symbol usually given to the c o m b i n e d potentials is U, so that U-
(I)g --[- ~ f
(2.17)
The final result for U is obtained from (2.13), the superposition of the f u n d a m e n t a l solutions of (2.14a, b, c), and from (2.16). By reason of symmetry, some of the terms in the superposition, referred to as "inadmissible terms," will vanish identically. Thus it can be shown that as a c o n s e q u e n c e of placing the origin of the spherical coordinate system at the center of mass of the earth, the terms with n -- 1, that is, with P1 a n d P1~, vanish. An additional s y m m e t r y comes about because the earth's spin axis, for dynamical reasons,
2.1
Gravity
29
coincides with high precision with the axis of m a x i m u m m o m e n t of inertia, hence with a principal axis. Although there is no exact rotational s y m m e t r y about this axis, it nevertheless m e a n s that two of the three products of inertia are zero, with the consequence that the term with P~ vanishes. Omitting the inadmissible terms in the summation, the final result for U can be written
{
u=Q
~-~(ae) n n=2
x (Cnm cos m~ +
JnPn(cos O) + ~ n=2
Pnm(COSO) m=l
Snrn sin rnA) } - 1f2202 sin 2 0
(2.18)
(with the proviso that in the double s u m m a t i o n the term with P2~ also vanishes). The factor used to nondimensionalize the distance from the mass center is the m e a n equatorial radius ae -- 6378.140 km. The first term in the main bracket comes from P0 and represents the gravitational potential of a spherically symmetric earth, hence is identical to the potential of a point mass. This term and the terms in the first s u m are i n d e p e n d e n t of the longitude. For this reason they are referred to as zonal spherical h a r m o n i c s . The terms in the double sum are referred to as tesseral spherical harmonics (Latin tessera: a piece of mosaic). Examples ofboth types are illustrated in Fig. 2.3, which shows the regions in which the spherical harmonics have positive versus negative values. The last term in (2.18) represents the potential of the centripetal acceleration. The constants In, Cnm, and Sm represent the mass distribution in the earth and are found by combining gravimetric, geodesic, and satellite measurements. The coefficient with the largest absolute value by far is 12, which derives from the equatorial bulge or flattening of the poles. Approximate values of the more important coefficients are listed in Table 2.1.
Figure 2.3 Examples of spherical harmonics: (a) zonal harmonic P4(cos 0); (b) tesseral harmonic p4(cos 0)(C4 cos4~ + S4 sin 4X).
30
C H A P T E R 2 Forces a n d Moments
Table 2.1 Spherical harmonics coefficients for the earth Tesseral h a r m o n i c s Zonal harmonics
n
m
Cm
Sm
]2 = -1082.70 10 -6 ]3 = 2.56 10 -6 ] 4 ~--1.58 10 -6 ]5 = 0.15 10 -6
2 3 3 3
2 1 2 3
1.57 10 -6 2.10 10 -6 0.25 10 -6 0.077 10 -6
-0.897 10 -6 0.16 10 -6 -0.27 10- 6
---
--0.59 10 -6
4
1
]7 --
0.44 10-6
4
2
0.074 10-6
4 4
3 4
0.053 10 -6 -0.0065 10-6
]6
0 . 1 7 3 10 -6
- 0 . 5 8 10-6
- 0 . 4 6 10 -6 0.16 lO-6
0.004 10-6 0.0023 10-6
Adapted from Ref. 2. Different sign c o n v e n t i o n s for the coefficients are s o m e t i m e s used, oft e n d e p e n d i n g on w h e t h e r the author's interest is in geodesics or in orbital m e c h a n i c s . D e p e n d i n g o n these c o n v e n t i o n s , the algebraic signs p r e c e d i n g the s u m m a t i o n s y m b o l s in (2.18) m a y differ. The z o n a l h a r m o n i c s have a greater effect on n e a r - e a r t h satellites t h a n the tesseral ones. The f o r m e r are n o t only larger b u t also p r o d u c e c u m u l a t i v e effects o n the satellite orbits, w h e r e a s the latter c a u s e oscillatory p e r t u r b a tions t h a t t e n d to m o r e nearly average out over r e p e a t e d orbits. Let u = -grad U
(2.19)
u(o, 0, ~.) is therefore the vector s u m of the gravitational a n d centripetal accelerations. Its radial, m e r i d i o n a l (in the direction o f i n c r e a s i n g colatitude, h e n c e positive f r o m n o r t h to south), a n d l o n g i t u d i n a l c o o r d i n a t e s are uo -
OU O0
10U '
uo -
0 O0
1
,
uz =
OU
0 s i n 0 O)~
(2.20)
h e n c e are o b t a i n e d by differentiation of (2.18). A n u m e r i c a l example, s u m m a r i z e d in Table 2.2, gives s o m e insight into the relative m a g n i t u d e s of the terms. The table applies to the p o i n t w h e r e a s p a c e c r a f t t h a t orbits the earth at an altitude of 300 k m crosses 45 ~ n o r t h latitude. The effect o f t h e e q u a t o r i a l bulge or flattening o f t h e e a r t h on the m e r i d ional gravitational c o m p o n e n t is evident. It is also of interest to n o t e t h a t the
Table 2.2 Acceleration corrections for the nonspherical earth for 0 = 45 ~ and 300 km altitude
Component
0 O k
Gravitational
Centripetal
acceleration
acceleration
+0.678 l O - 3 g o + 1.36 10 -3go O(lO-6)go
-1.728 lO-3go + 1.728 10 -3go 0
31
2.2 Thrust
centripetal acceleration is of the same order of magnitude as the radial and meridional gravitational corrections to the spherically symmetric earth.
2.2
Thrust In defining the thrust of a rocket motor as a force acting on the vehicle, one needs a definition that characterizes, to the extent possible, the motor performance in isolation from the other parts of the vehicle. When operating in the vacuum of space, the thrust is independent of the exterior of the vehicle. However, when operating in the lower atmosphere there will unavoidably be some interaction between the rocket plume and the air flow at the base of the vehicle. As is discussed further in Chap. 4, the gas flow in the nozzle and at its exit surface may then substantially deviate from the nominal velocity and pressure distribution present when operating in vacuum. Depending on the design of the rocket motor, flow separation and oblique shocks in the nozzle may occur, further altering the interaction of the rocket plume with the external air flow. One needs to distinguish therefore between the vacuum, or nominal, thrust of a rocket motor and the thrust at various altitudes in the atmosphere. An example of this variation of the thrust with altitude is shown in Fig. 2.4. The thrust of a rocket motor can be measured by the reaction force on load cells on a test stand. Small and medium-size motors are often tested in either a horizontal or vertical position in a low-pressure chamber. Large pumps, typically steam ejectors, are needed to compensated for the flow of the rocket gas entering the test chamber. Very large motors are usually tested in a vertical position, with the exhaust directed downward. They must be tested at ambient atmospheric pressure because otherwise the size and power requirement of the p u m p s would be so large as to be impractical. The corrections needed for determining the vacuum thrust must then be inferred from flight tests of similar
950 kN 290 s
255 s
0
T
10
20
Altitude (km)
30
40
50
F i g u r e 2.4 Altitude performance of the H-1 liquid-propellant rocket engine. From Ref. 5, Huzel D. K. et al., "Modern Engineering for the Design of Liquid Propellant Rocket Engines." Courtesy of the Rocketdyne Division of Rockwell International. Copyright 9 1992, AIAA--reprinted with permission.
32
C H A PT E R 2
Forcesand Moments
motors or must be estimated by analysis, often supported by tests at a reduced scale. In principle, the gas velocity and thermodynamic variables of state would not have to be exactly the same in a static test firing as compared with actual flight on an accelerating, and possibly spinning, vehicle. However, because the acceleration of the gas in the nozzle is m a n y orders of magnitude larger than the vehicle's acceleration, the difference arising from this cause between static tests and flight is negligible. In some solid-propellant motors, some slag may be retained in the motor case. Because the a m o u n t of slag may differ in flight from that measured in static test firings [6], the mass flow rates at the nozzle exit may differ slightly in the two modes of operation. However, in the majority of cases this effect can be neglected. IfdA designates the outward-directed unit vector, normal to the exit surface Aex of the rocket motor nozzle, u the gas velocity relative to the vehicle, and 0 the gas density, the mass flow rate through the nozzle is given by r~= f
dA 0 u . dA
(2.21)
ex
The change of m o m e n t u m exiting the nozzle per unit time, relative to the vehicle, is fA ~OU(U.dA)
(2.22)
ex
Ifm - m(t) is the instantaneous mass ofthe vehicle, V - V(t) the vehicle's center of mass velocity relative to inertial space, p the gas pressure at the nozzle exit plane, and p~ the ambient (i.e., atmospheric pressure; Pa - 0 in the vacuum of space), then conservation o f m o m e n t u m , applied to a control volume enclosing the vehicle and moving with it, results in the equation of motion
d(mV)/dt + fA (V + u)ou . d A - - /A ( p - pa) dA ex
(2.23)
ex
When the same motor is test fired on a test stand, since now V = 0, fA u Q u . d A - - f A ex
(p-pa)dA+T ex
where T is the reaction force exerted by the test stand on the vehicle. The measured thrust, Ft, is therefore -T, so that
Ft=-[/AexUQU'dA-t-~Aex(P--Pa)dA]
(2.24)
The first term on the right is often referred to as the velocity thrust. The second term represents the part of the thrust that can be ascribed to the pressure difference at the nozzle exit. It is referred to as the pressure thrust. When integrated over the entire vehicle surface, the force produced by the ambient pressure is, of course, zero; hence the term with Pa merely
2.2 Thrust
33
accounts for the fact that at the nozzle exit this pressure is replaced by the pressure of the propellant gas. In practice, the pressure thrust at high altitude or in vacuum is no more than a few percent of the velocity thrust. Nevertheless, in most cases it still needs to be considered because of the high sensitivity of the mass fraction available for payload as a function of the ratio of thrust to total propellant weight. At low altitude in the earth's atmosphere, depending on the nozzle design, the pressure thrust is often negative, that is, in a direction opposite to the m u c h larger velocity thrust. In place of computing the time rate of change of m o m e n t u m and the pressure at the nozzle exit, the same result for the thrust could also have been obtained by evaluating all pressure force components parallel to the thrust axis that act on the interior surfaces of the motor case and nozzle. This equality follows directly from the m o m e n t u m equation for a control volume b o u n d e d by the interior surfaces and by the nozzle exit surface. With this definition of the thrust, the m o m e n t u m equation (2.23) becomes d ( m V ) / d t + V/A 0U" d A -
Ft
-
0
ex
Noting that m -- - d i n / a t
(2.25)
and using (2.21), one obtains as the final expression of the equation of motion m d V / d t - Ft
(2.26)
The form of this equation is not quite as obvious as it might appear. It is important to note that the first term is m d V / d t , not d ( m V ) / d t . It may therefore be ofhistorical interest to mention here that Newton expressed his second law, albeit for constant mass, by the time derivative of the m o m e n t u m (his "vis motionis"), rather than of the velocity as is done in most elementary textbooks today. Perhaps ironically, Newton's formulation turns out to be correct relativistically (for constant rest mass), whereas the elementary textbook formulation would not hold.
2.2.1
Specific Impulse An important performance parameter of rocket motors is the specific impulse, Isp, defined by the equation
Ft
= g0/sp/~
(2.27)
where go - 9.80665 m/S 2 is the internationally agreed standard gravitational acceleration. The higher the specific impulse, the lower the rate ofpropellant consumption for a given thrust. As a consequence of the lower propellant mass needed for a given space mission, a high specific impulse will therefore result in a higher ratio of payload to propellant mass. For mainly historical reasons, the specific impulse is defined by the propellant weight flow rate g0r~, rather than by the mass flow rate r~ as
34
C H A PT E R
2
Forces a n d M o m e n t s
would have been more appropriate. Numerically, Isp is therefore expressed in seconds. The specific impulse that can be achieved in practice depends in part on the design of the rocket motor, particularly on the expansion ratio of the nozzle. But principally it depends on the chemical reaction energy of the propellant, o r - - in the case of nuclear-thermal motors on the reactor temperature. It depends relatively little on the size of the motor. Typical performance figures will be discussed in Chap. 4.
2.2.2
Tsiolkovsky's Rocket Equation In the absence of gravity and of all other forces except thrust, particularly simple conclusions can be drawn from (2.25) and (2.26). If, in addition, it is assumed that the thrust is in a direction tangential to the rocket's path, the motion becomes rectilinear and the equation of motion can be readily integrated. It follows that
fv~' dV =
-golsp
fm'~dm o
where mo and m~ are the initial and final mass, respectively, and similarly for the velocities. Carrying out the integration and solving for the mass ratio yields m = exp ( _ V1 - Vo) m0 g0Isp
(2.28)
emptymass,
In applications, the mass m~ would typically be the that is, the mass remaining after all the usable propellant has been consumed. It is also of interest to note that (2.28) holds independently of the time history of the mass flow rate and therefore of the thrust, as long as Isp is constant. Equation (2.28), in essentially this form, was first stated by Tsiolkovsky (Russian mathematics teacher; first publication on space travel 1895). Although actual rocket trajectories are almost always more complicated than a simple rectilinear path in a gravity-free space, the equation is important because it points out particularly clearly the need for a high specific impulse in space missions. These almost always require a large velocity increment. Because of its exponential dependence, the mass ratio, unless the specific impulse is high, can become very small, resulting in an unacceptably small payload mass for a given initial mass.
2.2.3
Example: The Sounding Rocket Sounding rockets are comparatively small vehicles, used most often for scientific purposes, particularly for the study of the earth's upper atmosphere and ionosphere. Sounding rockets therefore do not need to reach orbit. They are often launched along a near-vertical trajectory. We consider here a single-stage rocket, launched vertically. The mass flow rate r~ is assumed constant until the time when all propellant is exhausted (or until the propellant flow is deliberately terminated so as to
2.2
Thrust
35
control more precisely the final height attained). For simplicity of the calculation, the specific impulse Isp will be a s s u m e d to be i n d e p e n d e n t of the altitude h(t) and the aerodynamic drag will be neglected. The altitude reached by the rocket is a s s u m e d to be sufficiently small c o m p a r e d with the earth's radius that the gravitational acceleration can be a s s u m e d constant and equal to the standard value go. The guidance algorithm is p r o g r a m m e d to maintain, in an inertial reference, the thrust axis along the local vertical. The origin of time will be taken at the time of launch. The time at thrust termination will be designated by t~. The mass of the vehicle is re(t), with m0 at launch and ml at, and after, thrust termination. During the powered phase, from (2.27) and (2.24),
mCt) d Wdt
(0 < t < t~)
gom(t)
= go/sp/'~--
(2.29)
where tl = m ~
=
fit
(1_
m l ) Isp
mo
#o
(2.30)
and where Ft fl0 = g0m0'
fl0 > 1
(2.30')
is the ratio of thrust to initial weight. Integrated, with V = 0 at t = 0 and m(t) = mo - ~t, the velocity of the rocket becomes
(
V = g0Ispln 1
1 ) - got
(0 <_t <_tl)
fl0t/)sp
(2.31)
The velocity is seen to reach a m a x i m u m V1 = g0Isp ln(mo/mx) - g0(1 -
mx/mo)Isp/flo
(2.31')
at the time tl of thrust termination. The second term on the right in (2.31') reflects the fact that the m a x i m u m velocity achievable for a given mass ratio and specific impulse becomes smaller for smaller values of the ratio of thrust to initial weight. In the limit as fl0 ~ c~, that is, if an infinite thrust could be applied instantaneously at the time of launch, the second term would vanish. Comparison with the general case then suggests referring to g0(1 -
ml/mo)Isp/flo
as the gravity loss term. This loss becomes i m p o r t a n t for values of flo close to 1. Especially in the case of large vehicles, because of the physical limitations imposed on the size of the rocket motor and on the propellant feed rate, the designer is often forced to accept a substantial loss of this type. Integrating a second time, with h - 0 at t - 0, results in
h golsp[Isp-( Isp O ( Lfl0
fl-0-0-
1-1n
(
fl~ 1-/-~p
gOt2 2
(0<_ It_< tl) (2.32)
36
C H A P T E R 2 Forcesand Moments for the height as function of time during the powered phase. At thrust termination, when t = tl, the height becomes
hl
goI2sp[ fig
ml(
fl0-fl0~
too) l+lnm11
-2
1(ml) 2] 1---mo
(2.32')
Beyond this height, the rocket is coasting, with a velocity
V = V~ - go(t- ta)
(t > tl)
(2.33)
At peak height, V = 0. Hence, from (2.31'), the time after the launch needed to reach peak height is t2 = Isp ln(m0/m~)
(2.34)
provided, of course, that Ft > g0mo. To obtain the peak height, h2, we observe that from (2.33) integration from t~ to t2 and substitution of 111 and t2 gives h2 = ha + 2fl02 1 - --m0 - fl0 In --m~
(2.35)
where hi is obtained from (2.32'). Figure 2.5 illustrates the rocket's velocity as a function of time for three values of the ratio of thrust to initial mass. The ascending branches in the figure correspond to the powered phase, as given by (2.31), and are independent of the mass ratio. The descending branch in the nondimensional representation in the figure has a slope o f - 1 and corresponds to the coasting phase. It is independent of the ratio of thrust m initial weight but depends on the mass ratio. For a numerical example, we assume a specific impulse of 250 s (which is representative for a sounding rocket solid-propellant motor, with altitude-
.4~
,,,,,
t .3
oo 1
~Thrust goT-no
..V go/sp .2
Termination
/39o
.1
0
I
.1
.
t/Isp Figure 2.5 Velocityversus time ofavertically ascending sounding rocket, for a mass ratio of 1.60.
2.3 Aerodynamic Forces and Moments
37
averaged performance), a mass ratio of 1.60, a thrust of 10,000 N, and an initial mass of 500 kg. From the relations derived before this, one then finds that the time to burnout (thrust termination) is 46 s, at which time the altitude is 14.1 km and the velocity 701 m/s. Peak height is reached at 117 s after launch, at a height of 39.2 km.
2.3
Aerodynamic Forces and Moments Aerodynamic forces and moments need to be considered for the a s c e n t of launch vehicles in the earth's or other planetary atmospheres. To avoid large structural bending m o m e n t s ~ for which they are not d e s i g n e d ~ major launch vehicles are almost always launched vertically. Starting from low speed, where the fluid mechanics of incompressible fluids applies, they reach hypersonic speeds in the upper atmosphere, where the path may then deviate from the vertical. The drag experienced by these vehicles is much smaller than the thrust. Preflight predictions of the drag do not require the same accuracy as is needed in the performance prediction of aircraft. Also, the aerodynamic shape of launch vehicles does not need to conform to the exacting demand for low drag contours familiar from aircraft designs. More important, in the case of large launch vehicles, is the reduction of the structural weight by choosing, for instance, simple conical and cylindrical shapes without much fairing or other drag reduction features. Smaller vehicles may be air launched. They are then usually designed for aerodynamic lift, hence must sustain the ensuing bending moments. Extremely demanding is the aerodynamic design of vehicles designed for reentryinto the earth's atmosphere or, more generally, for aerobraking in a planetary atmosphere. In addition to lift, drag, and aerodynamic moments, protection of the vehicle's structure against excessive heating then becomes of paramount importance. Some space missions postulate spacecraft that are designed to lose speed by repeatedly entering and then exiting the atmosphere of planets. They may do this by a sequence of skipping in and out of the atmosphere, followed by a segment of an orbit in space. This technique was first demonstrated in the Venusian atmosphere by the United States Magellan Venus Orbiter. This spacecraft had neither an aerodynamic shroud nor special heat protection, but it could be successfully controlled by its thrusters to avoid aerodynamic instabilities and overheating. Aerobraking of a spacecraft may also be used for reducing speed prior to the launch of a scientific probe designed to descend to the planet's surface or for changing the orbital parameters of a spacecraft without intent of landing.
2.3.1
Ascent of Launch Vehicles
The reader is likely to be aware of the existence of a very large body of aerodynamic literature, both theoretical and experimental. Much of this American and European literature is contained in monographs published by NASA, ESA (the European Space Agency) and AGARD (the North Atlantic
38
C H A P T E R 2 Forces a n d M o m e n t s Treaty Organization), as well as in textbooks, for example, Ref. 7. In the following we confine ourself to a brief, qualitative discussion. Launch vehicles usually have a roughly cylindrical, elongated shape. Lift and drag coefficients for such shapes have been measured in wind tunnels. Because of the inevitable, practical limitations ofwind tunnels, the Reynolds and Mach numbers at which such tests can be conducted are m u c h smaller than what would be required for a complete aerodynamic characterization of the ascent of launch vehicles. Extrapolations based on theoretical considerations must therefore be made. But because the increase in propellant consumption caused by aerodynamic drag is relatively small, great accuracy, comparable to that needed for aircraft, is not required. There also exist analytic solutions for high-speed, inviscid flows about slender, cylindrical bodies [8]. They can provide approximate information on the lift, aerodynamic moment, and supersonic wave drag. At the cost of extensive programming and calculating on high-speed computers, more realistic results can be obtained by numerical methods ("computational aerodynamics"). These method allow an accurate and detailed representation of the vehicle's outer surface and of the resulting potential flow. More subject to errors is the modeling of turbulent boundary layers, of the laminar-to-turbulent transition, and of flow separation. To illustrate, Fig. 2.6 from Ref. 9 shows the pressure distribution at transonic speed (M = 1.1) for the Space Shuttle Orbiter in its plane of symmetry on the top side. The pressure, p, is expressed by the "pressure coefficient,"
o~= -3~ m
-0.5
_
Cp o
i
0.5
0
0.25
0.50
0.75
1.00
Figure 2.6 Pressure coefficient, Space Shuttle Orbiter upper side. Solid and dashed lines, computed; (zx) wind tunnel data; (o) flight data. Angle of attack, -3~ Mach number, 1.10. (From Ref. 9 Martin, E W. and Slotnick, ]. P., "Flow Computation for the Space Shuttle in Ascent Mode", in Progress in Astronautics and Aeronautics, Seabass, A. R., ed. Copyright 9 1990, AIAA-- reprinted with permission.
39
2.3 Aerodynamic Forces and Moments
that is, by the quantity (p - Pa) / [(0a/2) V2] shown as a function of the body length (Pa - ambient atmospheric pressure, ~Oa = ambient density, and V = flight velocity). The solid and dashed lines are the result of two different numerical calculations. The triangles represent wind tunnel data. Actual flight data are indicated by the two circles. The importance of the lift and of the m o m e n t derives largely from the need to consider the effect of u p p e r a t m o s p h e r e winds. Moderate winds at the launch pad are less dangerous to the integrity of the vehicle than those at higher altitude. Not only can the wind speeds be m u c h higher, but also the vehicle velocity is high. Wind shear, in particular, necessitates rapid and forceful steering of the vehicle, imposing large bending m o m e n t s on the structure. For this reason, prior to the launch of a major vehicle, the wind velocity above the launch site is measured. The required data can be obtained by radar and by balloons. Because the vehicle's velocity is still low, the aerodynamic loads in the absence of winds are relatively small at low altitude. At high altitude, they are also small because of the low density. Between these extremes there is a density and velocity condition where the loads reach a maximum. A suitable measure for the severity of these loads is the stagnation pressure. Figure 2.7 shows the ratio of stagnation pressure to sea level ambient
M = 17.6
40
1.5
30
9.19
q
E
v
gomo
7.16
t-.. v
3.50 5.25
-o
+~ m
20
7.00 1.25 0.78
10
1.38
0.37
I ~ 0
2.03
0.92
I 1.0
I 2.0
I 3.0
1/2oav2 Psi
Figure 2.7 Vertical ascent of a launch vehicle: ratio of stagnation pressure to sea level pressure, psi, as a function of altitude, for three values ofthrust to gross mass. Isp, 350 s; M, Mach number.
40
C HA P T E R 2 Forces a n d Moments pressure as a function of altitude. The curves are for three values of the thrust relative to the launch vehicle's initial weight. A specific impulse of 350 s is assumed. Also indicated are the Mach numbers. As this figure shows, the peak of the stagnation pressure occurs around 10 km altitude, almost independent of the thrust-to-weight ratio. At this altitude, strong buffeting may occur, particularly w h e n there is flow separation from parts of the vehicle that are not shaped aerodynamically. Buffeting can be dangerous because of the vibration loads imposed on the vehicle. For this reason, on some vehicles with engines that can be throttled, the thrust is programmed so that it is decreased in passing the region of high stagnation pressure, before increasing again as the vehicle reaches higher altitudes.
2.4
Free Molecule Flow With increasing altitude above the earth's or planetary surface, the aerodynamic characteristics ~ particularly lift, drag, and convective heat transf e r ~ undergo fundamental changes. At one extreme there is classical aerodynamics, which is based on continuum mechanics. In space technology, this is applicable primarily to the takeoff and flight of launch vehicles in the lower atmosphere. At the other extreme is the regime of free molecule flow, applicable in particular to satellites in low earth orbits. Classical aerodynamics applies when the m e a n free path between successive collisions of the gas molecules is very small in comparison with the thickness of the boundary layer (hence, a f o r t i o r i when compared with the body dimensions). On the contrary, in free molecule flows the mean free path is m u c h larger than the principal dimensions of the body that is immersed in the flow. For instance, at an altitude of 300 km in the earth's atmosphere, where a n u m b e r of low-earth-orbiting satellites operate, the length of the m e a n free path is about 900 m. There exists a continuity of different flow regimes from the continuum mechanics of classical aerodynamics to the free molecule flow. The two intermediate regimes that are distinguished are referred to as "slip flow," typical for slightly rarefied gas flows, and "transition flow," typical for moderately rarefied gas flows. The theoretical treatment of these intermediate flow regimes is difficult and incomplete and will not be considered here. Historically, free molecule flows were first studied in laboratory systems [10, 11]. The theoretical and experimental results obtained have since been applied to the study of the hypersonic flight of space vehicles in the rarefied upper atmospheres, in particular of Earth and of Mars. Other than depending on the nature of the gas or gas mixture and on the density, the m e a n free path is also somewhat, but only weakly, dependent on the energy of collision of the molecules. Experimental data are usually stated for 760 m m Hg pressure and 15~ temperature and are expressed in terms ofthe effective collision diameter, D, ofthe molecules. The conversion of these data to the conditions of a rarefied atmosphere by the ideal gas law is straightforward. If N is the n u m b e r of molecules per unit volume and if the molecules are assumed to be elastic spheres with a diameter D and a
2.4 Free Molecule Flow
41
Table 2.3 Molecular collision diameters and mean free paths at normal temperature and pressure
Gas 02 N2
CO2 He
Molecular weight
Effective diameter (cm)
Mean free path at 760 m m Hg (cm)
32.00 28.02 44.00 4.002
3.61 10-8 3.75 10-8 4.59 10-8 2.18 10-8
6.79 10 -6 6.28 10-6 4.19 10 -6 18.62 10 -6
Maxwellian velocity distribution, it can be shown that the average length, )~, of the free paths is given by X =
1 ~/27r N D 2
(2.36)
The ratio of the m e a n free path to the characteristic vehicle dimension, l (such as the length or effective diameter of a reentry probe or spacecraft), is known as the Knudsen n u m b e r K = ~./l. In Table 2.3 the effective diameters for intermolecular collisions and the m e a n free paths at normal temperature and pressure are listed for several gases. Carbon dioxide is the principal constituent ofthe atmosphere of Mars. Helium becomes important in the earth's atmosphere above 400 km. Currently available data on properties such as density and temperature in the upper atmosphere (the t h e r m o s p h e r e and the exosphere) of the earth are subject to considerable uncertainties. To some extent, the data d e p e n d on the type of experimental m e t h o d that is being used. There are strong variations that d e p e n d on the solar flux in the extreme ultraviolet and hence are synchronous with the 11-year solar cycle. In addition to latitude and seasonal variations, there are also variations with local time: on the sunlit side of the earth, the atmosphere is heated and rises, with the consequence that the density in the u p p e r atmosphere increases. The opposite applies to the night side. For low-earth-orbiting satellites these 12-hour variations tend to average out. They are important, however, for predicting the trajectories of reentering spacecraft and for estimating the effects of aerobraking maneuvers in the u p p e r atmosphere. In the United States, several models for the properties of the earth's upper atmosphere are in current use. Among t h e m is the U.S. Standard A t m o s p h e r e 1976 and the newer MSIS-86 Model (Mass Spectrometer Incoherent Scatter model). The latter is based on in situ data from satellites and rocket probes and also on m e a s u r e m e n t s from incoherent scatter stations on the ground. Figure 2.8 indicates average n u m b e r densities (molecules per unit volume) of the electrically neutral species in t h e earth's t h e r m o s p h e r e [12]. The species, roughly in order of their importance, are atomic oxygen, helium, atomic hydrogen, molecular nitrogen and oxygen, and argon. More recently, atomic nitrogen has also been included in the data.
42
C H A P T E R 2 Forces a n d M o m e n t s 1000 900
N t l -\I "\,,He
800 700
v
_"
600
- -
500
..
400
3OO 2OO 100
I
I
I
I
',
I
I
\O
I
I
I
I
I
I
I
I
06 107 108 109 1010 1011 1012 1013 10 TM 1015 1016 1017 10181019 1020 Number density (m 3)
Figure 2.8 Average number densities ofthe electrically neutral species in the earth's upper atmosphere. (From Ref. 12.) Copyright 9 1995 Princeton University Press. By permission. Figure 2.9 [13] shows averaged data for the combined (mass-weighted) density of these species and for the temperature at 400 km altitude. The data are shown as functions of the solar activity and include three different models. Making use of a result of the kinetic theory of gases, the m e a n free path can be related to the kinematic viscosity and therefore to the Reynolds number. This makes it possible to express the Knudsen n u m b e r in a form that is particularly convenient for applications in fluid mechanics. The result is
K = 1.26~/~M/Re
(2.37)
where M is the Mach number, Re the Reynolds number, and y the ratio of the specific heats. Here, K and Re are based on the same characteristic body length I. Free molecule flow is usually defined as the region for which M / R e > 3.
5x10-14 f
,
I
,
I
,
I
,...- .- .. t 1400
1200 crl
E
10-14
U
E v
1000
cn
c~
E
r-
C3
10-15 5xl 0-16 70
~
/
~ J
Density
I 110
t
I 150 Solar Flux
....
J77
.......
MSIS-83
I
I 190
8O0
230
600
Figure 2.9 Average density and temperature ofthe electrically neutral earth atmosphere at 400 km altitude, as functions of the solar activity index. For three different models. (From Ref. 13.)
2.4 FreeMolecule Flow
43
If, as will now be assumed, the m e a n free path is large compared with the characteristic dimensions of the spacecraft, the paths of the gas molecules prior to their impact on a spacecraft surface are no longer influenced by the presence of the body. Neither shock fronts nor boundary layers are formed. In applications to orbiting spacecraft in the upper atmosphere, the relative velocity of spacecraft and atmosphere is hypersonic. In this case, the thermal motion of the incident molecules can be neglected. Also, it will be assumed that these molecules reach the surface without having first been reflected from another spacecraft surface. The incident molecules therefore travel on parallel, rectilinear paths, directly incident on the surface. In turn, the molecules are reflected from the surface. This process can be described approximately as being a mixture of diffuse and specular reflections. The reflected molecules will collide with the incoming stream, as well as among themselves, but only at a large distance from the spacecraft so that they do not react back on the surface. The diffusely reflected molecules, through their multiple interactions with the surface before being reemitted, are at least partially accommodated to the temperature of the surface. Their velocity distribution is approximately in accordance with the Maxwell-Boltzmann distribution
~3/2 mv2~ 47r(27rkT/ v2exp(-2kTj dv m
f(v) dv
(2.38)
where f ( v ) d v is the probability that a molecule has a velocity between v and v + dv and m is the mass of the molecule, k the Boltzmann constant = 1.3806 10 -23 kg m 2/(s2IO, and T the temperature. It follows by integrating (2.38) that the average velocity, ~, and the root m e a n square velocity, v, are
~-
-~---~/
,
~=
-
-
(2.39)
The average translational kinetic energy per molecule due to the thermal motion is therefore 3 kT. Figure 2.10 shows schematically the incident flow of molecules on a surface element and the reflected flow. For hypersonic flight, the thermal motion of the incident molecules can be neglected. The angle of incidence is designated by ~i, the incident speed by V, and the temperature of the surface (wall) by Tw. The reflected flow can be approximately represented by the sum of two parts: a specular reflection at speed V and angle -qh, and a diffuse reflection with a Maxwellian velocity distribution at temperature Tr into the half-space above the surface element. The specular part represents elastic collisions with the surface, whereas the diffuse part represents multiple interactions with surface atoms, sufficient to establish approximately a thermal equilibrium among the molecules before they scatter in r a n d o m directions. The combination of incident and reflected flows imparts m o m e n t u m and energy to the surface. The number, say I, ofmolecules of a specified kind (for instance, atomic oxygen) striking the surface and being reflected, per unit time and unit area
44
C H A P T E R 2 Forces a n d M o m e n t s
Incident Flow (V)
)
Diffuse Reflection (
Figure 2.10
Reflection of a free molecule flow from a surface.
of the surface, is given by
I = N V cos ~i
(2.40)
The fraction of the incident molecules reflected diffusely is usually given the symbol a. Depending on the nature of the surface and of the incident molecules, a typically ranges from 0.8 to 1. The recoil of these molecules produces a force normal to the surface but, by symmetry, not a tangential force. The fraction of molecules reflected specularly is 1 - a. Their recoil causes both a normal and a tangential force. A second, nondimensional coefficient, the Smoluchowski-Knudsen acc o m m o d a t i o n coefficient, a, is introduced, which is defined by the equation c~ =
Ei-
Erd
Ei-
Ew
(2.41)
where Ei is the energy flux (energy per unit time and unit area of the surface) of the incident molecules, Erd the energy flux of the diffusely reflected molecules, and Ew the energy flux of these molecules if on the average they all had assumed the surface (wall) temperature. This coefficient is therefore a measure of the degree to which the energy of the diffusely reflected molecules has been "accommodated" to the surface temperature. For perfect accommodation, Erd ---- Ew, therefore a = 1. Depending on the types of surface and incident molecules, a generally ranges from 0.8 to 1.0. Evidently, =2 Ei =
I ~,mV2 2
Erd -- ~ I ~ , mrdv 2
Ew = a I _3k Tw 2
(2.42)
Since the Smoluchowski-Knudsen coefficient refers to the energy rather than to the m o m e n t u m , a slight conversion is necessary for calculating the force exerted by the molecules on the surface. Solving (2.41) for Erd and
2.4 Free Molecule Flow
45
making the substitutions indicated by (2.42), =2
a From
(2.39),
mVrd
2
=2 -- ~ 3 7t'Vrd -2 , l)rd
Vrd--
= (1 - o r )
mV 2 3 2 + a a -~ kTw
(2.43)
therefore
4~ 1 [(1-c~) ~
O"
V2
3k~w 1
- 2 - -}- ~ i ~
(2.44)
The free molecule pressure, Pfm, exerted on the surface is the sum of the pressures from the incident flow, from the recoil of the specularly reflected flow, and from the recoil of the diffusely reflected flow. Hence Pfm -- Pi + Prs + Prd
where Pi = I m V c o s ~bi and Prs = (1 - a ) I m V c o s ~bi. The pressure resulting fro m the diffusely reflected flow is obtained by integrating the surface no rmal c o m p o n e n t of the velocity V~d over the unit hemisphere above the surface element (Fig. 2.10). Hence Prd -- O"ImVrd f f
COS ~
dfZ = zccr I m v r d
The final result for the pressure exerted by the molecules on the surface, obtained by s u m m i n g the three terms, is Pfm
-
-
I m[ (2 - a) V c o s ~i + 7Z"O"Urd]
(2.45a)
Similarly, the shear stress, Sfm, is Sfm = Si + Srs + Srd
where Si -'- I m V sin ~bi, Srs -- -- (1 - a ) I m V sin ~i, and, by symmetry, Srd = 0. Therefore Sfm --- O"I m V s i n ~i
(2.45b)
In general, there will be several species of molecules in the flow, with different masses, and different a's and c~'s. Because in free molecule flow the molecules do not interact, the normal and tangential forces on the surface can be obtained simply by s u m m i n g the contributions from the various species. Depending on the geometry ofthe external surfaces, these pressures and shear stresses acting on the surface of a spacecraft will produce drag, lift, and a moment. Drag and lift are important in aerobraking into a planetary atmosphere and in estimating the remaining life of a spacecraft in the upper atmosphere before it reenters. Because of its effect on spacecraft attitude control, the m o m e n t also can be important. In the earth's atmosphere, at 400 km altitude, the flee-molecular torque on a medium-size spacecraft can be of the order of 10 -4 N m, which is larger than the solar radiation torque. At altitudes above 1000 km, it will tend to be smaller than either the solar radiation or gravity torques.
46
C H A P T E R 2 ForcesandMoments 300l
I 28~F
I
I
I
I
I
I
I
I
I
[ 9
220I 200F 180[
0
4
8
12 16 20 24 28 32 Decrease in altitude in one day (km)
36
40
Figure 2.11 U.S. Space Shuttle Orbiter: decrease in altitude per day at no thrust. For the three basic attitudes. (Adapted from Ref. 14.) It has been found that the upper atmosphere largely corotates with the earth. The motion in a nonrotating reference frame is therefore from west to east along circles of constant latitude, approximately with a velocity Va = 2Jrrg sin O/Pg (rg - earth radius, 0 = colatitude, and the period Pg - 24 h). At the equator, Va - 0.465 km/s, which is small compared with the orbital velocity Vcrc of about 7.75 k m / s of low-earth-orbiting satellites (Sect. 3.7). The velocity, V, of the spacecraft relative to the atmosphere, that is, the velocity of the incident molecules in hypersonic flight, is therefore the vector difference of Vcrc and Va. As is readily shown, the angle 0 between the two vectors is given by cos 0 - cos ieq/sin 0, where ieq is the inclination of the orbit relative to the equator. As an example of the aerodynamic forces in the earth's upper atmosphere, Fig. 2.11 shows the rate of orbital decay (at zero thrust) of the U.S. Space Shuttle Orbiter. At the range of altitudes shown, the atmospheric density decreases by about a factor of 10 for each 100-km increase in altitude. Free molecule flow in the case of the Orbiter applies down to about 280 km. Below this altitude, the regions of "transition flow" and "slip flow" apply. Drag-free satellites in low earth orbits are of interest in geodetic research. Their proximity to the earth's surface makes their trajectories sensitive to the earth's mass distribution. Gravitational disturbances of the order of 10 -12 times normal gravity have been detected. Spacecraft of this type use an internal, flee-floating, proofmass, shielded from the atmosphere by an external shell. Small thrusters on the shell are controlled such that the relative position of the proof mass to the shell remains approximately constant. The shell therefore is constrained to follow the proof mass.
2.5 Solar Radiation Pressure
47
The method used to calculate the pressure and shear stress also lends itself to an estimate of the heat transfer in free molecule flows. The energy flux incident on the surface is Ei = (1/2)ImV2; the energy flux leaving the surface is E r . The a m o u n t of heat transferred into the surface per unit time and unit area is q -- Ei - E r . Therefore, from the definition (2.41) of the accommodation coefficient, assuming a -- 1, q
-
3 kTw) aI(-~1 m V 2 - -2
In the general case in which several molecular species Sj, j = 1, 2, . . . , are present, q
2.5
_
~ [ o l j l j ( 8 9 mj V 2 J
3 kTw)]
(2.46)
Solar Radiation Pressure Solar radiation can exert an appreciable force and torque on a spacecraft. For instance, at an altitude of 500 km in the earth's atmosphere, the effects of solar radiation on an orbiting spacecraft are of the same order of magnitude as the atmospheric drag. At higher altitudes, the solar radiation pressure predominates. The radiation pressure decreases with the inverse square of the distance from the sun. (Solar radiation needs to be distinguished from the solar wind, which is the n a m e given to the particle stream, mostly protons and electrons, that emanates from the sun. The pressure produced by the solar wind is negligible by comparison with the radiation pressure.) The principal interest in solar radiation pressure derives from the disturbing effects it produces over long periods of time on the trajectory and attitude of spacecraft, on earth satellites, as well as on spacecraft on missions in the larger solar system. Although its practicality has not been proved as yet, m a n y authors have proposed to utilize the radiation pressure for solar sailing, particularly for small, scientific payloads to the inner planets or near-earth asteroids. The advantage, of course, is that no propellant needs to be carried. As has been shown theoretically, trajectory changescan be accomplished with good control. For missions for which low thrust is acceptable, solar sails must compete with the more versatile means that are provided by electric propulsion, such as ion propulsion. The m o m e n t u m of a photon is E / c (E = energy, c = velocity of light in vacuum). The energy flux (energy per unit area and unit time) of the solar photons in the earth's vicinity (i.e., at 1 AU) has been measured repeatedly. Its value, 1353 W / m 2, is referred to as the solar constant, designated here by the symbol jh. It follows that the corresponding m o m e n t u m flux that is incident on a surface perpendicular to the rays is jh/C = 4.51 10 -6 N / m 2. If the surface is nonabsorbing and specularly reflecting, as would be the case, for instance, for solar sails, the pressure on the surface is double this value because of the recoil of the photons. For instance, at the average solar
48
C H A PT E R 2 Forcesand Moments
distance of Mercury, the pressure on a unit area perpendicular to the rays would be 5.9 10 -5 N / m a. The energy, ]h, of the photons, incident at an angle ~i (measured from the surface normal) on a surface of unit area, per unit time is ]h -- Jh(rg/r2) COS~bi
(2.47a)
where r is the distance from the sun and rg = 1 ALl. It follows that the corresponding flux of m o m e n t u m is
]h/C
-
-
(Jh/c)(rg/r 2) COS~i
(2.47b)
In what follows, we consider two separate cases. In the first, the surface is assumed to be nonabsorbing and reflecting specularly. In the second, the surface is partially absorbing and reflects the remainder diffusely, following Lambert's empirical law for such surfaces. Nonabsorbing and reflecting specularly: The total pressure, Ph, is
Ph -- 2(jh/c)(rg/r2)
COS2 t~i
(2.48a)
Because of the cancellation of the effect of the incident by the reflected photons, the shear stress Sh = 0
(2.48b)
Partially absorbing with diffuse reflection: The absorption of light by a surface is characterized by the absorption coefficient ah for solar photons. It should be noted that, as usually defined, this coefficient refers to the absorption of energy, not of m o m e n t u m . The conversion, however is simple: The reflection of light by partially absorbing surfaces usually follows closely Lambert's law (Lambert, 1728-1777). This law states that the intensity of the reflection falls off with the cosine of the angle, say ~, from the surface normal. (It may be noted here that the diffuse reflection of photons differs from that of molecules, discussed in the preceding section. In the latter case, thermalization is assumed, with the result that the intensity of the flux is independent of the angle.) The flux of reflected energy is therefore (1 -- C~h)Jh = i n / / C O S
~b d ~
where in is the intensity (energy per unit area and unit time and per unit solid angle) that is reflected in the direction of the surface normal. The integration is over all solid angles ~2, that is, over the surface of the unit hemisphere above the reflecting surface element. Carrying out the integration and solving for in gives in = 7r -1 (1 -- Cth)]h
The pressure, Phr, that results from the reflected photons is obtained from Phr
= (in/C) f f
COS2 ~b dQ
Integrating over the solid angles as before and substituting for in and for ]h
2.6
Atmospheric Entry
49
from (2.47a) gives Phr = 2(3c)-'(1 - O~h)jh(rg/r 2) c o s ~i
Finally, adding to this the pressure from the incident photons then gives the total pressure from solar radiation,
Ph --Cjh/C)(rgl
COS+ [ Cl
cos
(2.49a)
By symmetry, the shear stress that results from the reflected photons vanishes. Therefore the total shear stress, Sh, equals that produced by the incident photons. Hence
Sh
-
(jh/c)(rglr 2) cos~i sin ~i
(2.49b)
2.6 Atmospheric Entry Reentry into the earth's atmosphere or entry into a planetary atmosphere calls for an understanding of the aerodynamic forces and moments on the vehicle and of the heat transferred to it. The flight paths are determined by gravity, drag, and lift. The lift, and to a lesser extent the drag, can be modulated by changes ofthe angle ofattack, either by small thrusters or by movable aerodynamic surfaces. Also, the trajectories do not need to be restricted to a plane: by banking the vehicle, without thrust, a limited amount of cross-range can be obtained. The purpose, most often, is to steer toward an intended touchdown point. Banking can also be used to slow a vehicle that might otherwise overshoot the touchdown point. If the flight path is meant to be in a plane and no active control is intended, care needs to be taken so that small asymmetries of the vehicle ~ even ifit is supposedly axisymmetric ~ do not produce forces perpendicular to the plane. By imparting a slow rotation to the vehicle, the effect produced by such side forces can be eliminated. Flight path control of manned vehicles is particularly demanding. Not only is high accuracy required in steering toward the landing point, but also the maximum deceleration may not exceed the physiologically set limit. A major uncertainty affecting flight paths is caused by the large daily, seasonal, and solar cycle-induced variability of the atmospheric density at high altitudes. Although the density there is low, the high entry speed produces appreciable drag and lift, which can be of a magnitude comparable to those at lower altitudes, where the density is much higher but also the velocity much lower. Entry at orbital velocities into a planetary atmosphere causes intense heating. The difficulties in predicting and guarding against the heat transferred to the vehicle are very great compared with the more conventional heat transfer problems encountered in other engineering applications. At speeds comparable to orbital velocities, the air (or carbon dioxide in the case of Mars) is partially dissociated behind the gasdynamic shocks and in the boundary layers. The various chemical species that are generated will diffuse in the boundary layers and can recombine at the surface, giving up their heat of reassociation to the surface. Because the ionization potentials are typically higher than those for dissociation, the degree of ionization is
50
C H A P T E R 2 Forces a n d M o m e n t s
weak by comparison, but often sufficient to cause a microwave communications blackout. Depending on the severity of the temperature rise, there can also be heat transfer by radiation from the gas, in addition to the convective heating. Much research has been conducted to produce heat protection materials. These materials, for instance lightweight, foamed ceramics, must have low heat conductivity, yet sufficient mechanical strength to withstand the aerodynamic pressure and shear. Ablation of the heat protection materials prevents their use on more than a limited n u m b e r of missions but has the advantage of providing a heat barrier against the hot gas by outgassing. If the surface is black, a major part of the heat incident on these materials is radiated into space. In addition to depending on aerodynamic studies of heat and mass transfer, predictions of the heating and its effect on the materials depend much on studies in chemical kinetics. This field, which is aptly called aerot h e r m o c h e m i s t r y is beyond the scope of this book. The interested reader is referred to Refs. 15 and 16 at the end of this chapter. For illustrative purposes, calculated surface temperatures at reentry of the Space Shuttle Orbiter are indicated in Fig. 2.12 from Ref. 17. Also indicated are the critical temperatures that are reached at maximum yaw of the Orbiter during ascent.
1260~
Lower Surface
109o~ ~ ~26o~ . " ,,'
1090~
~
~
0 0oc \ ><" ,'- --
: ~.-.."-: -
~
1410~ / - . t ' - I 1100~
. - 4-~ ..,-_. . .... -,'---,1~260o0 : . ~ ~44ooc
-':_. ....
"
.-,'"
,-~-,
,
....
osooc
. ..... "" "~: -~-[ 2 -
Jl 1
"950~
..... o~u~L., 430~
400~ ~ .
141ooc~
.
.
.
.
430~ 370~176 aonor 650~ "" ~ 400~176 "-" " ~ L'" 'Y 450o0/.480o0~~~//,'~,_-_~. .
.
.
.
*540~ / /~ ,~. "1180~ ",1220o 0
.
",_-.:_:,---.-..-:. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
, \ 980~ 1100~
oc,.4oooc oC/-440oC
~ooc
Figure 2.12 Calculated isotherms, Space Shuttle Orbiter, at reentry. (* denotes ascent temperatures at a yaw of 8.0~ From Ref. 17, Peake, D. ]. and Tobak, M. The original version of this material was first published by the Advisory Group for Aerospace Research and Development, North Atlantic Treaty Organization (AGARD/NATO) in 'IAGARDOgraph AG-252 __ ThreeDimensional Interactions and Vortical Flows with Emphasis on High Speeds," ]uly 1980, by permission.
2.6 Atmospheric Entry
v
51
Start of Entry
/
Figure 2 . 1 3 Entry into a planetary atmosphere, schematic. Similarly, the materials research, much of it being conducted in arc jets, cannot be treated here. A good summary of the calculations of flight paths and estimates of the heat transfer is contained in Chap. 6 of Ref. 14. In what follows, we illustrate the calculation of a reentry flight into the earth's atmosphere. This calculation, to stay simple, makes a number of assumptions that are only qualitative. The calculation, however, is sufficient to provide insight into the general problem. As indicated schematically in Fig. 2.13, a flight path in the plane of the figure is assumed. The altitude above sea level is designated by h. The vehicle is supposed to have drag, designated by D(h), but no lift or thrust. A constant drag coefficient, CD, will be assumed. The velocity vector, V(h), can be taken as the velocity in inertial space, but the difference between it and the vehicle's velocity relative to the atmosphere will be neglected. This assumption is quite crude: if the entry velocity, taken conventionally at about 120 km altitude, is the low-earth-orbit velocity, the vehicle velocity would be about 7.75 km/s, corresponding to a Mach number of about 20. But below this altitude, the drag rapidly increases, greatly reducing the vehicle's velocity. By contrast, the earth's, hence the atmosphere's, rotational velocity at the equator is 0.465 km/s, corresponding to a Mach number of about 1.35 at sea level. The flight angle, y (h), that is, the angle between the velocity and the local horizontal, is taken positive for a downward path. Also let m the mass of the vehicle, r(h) its distance from the earth's center, and rg the earth's radius. The change with altitude of the gravitational acceleration, g, will be neglected. The density o(h) is assumed to be an exponential function ofthe altitude. This can be roughly justified by assuming that the atmosphere satisfies the ideal gas law, at a temperature assumed to be independent of the altitude. From the hydrostatic equation, d p = - g o dh (p = atmospheric pressure), it then follows that R T do = - g o dh
(2.50)
Integration between the altitudes hi and h2 and the corresponding densities 01 and 02 results in h2 - hi = ( R T / g ) ln(01/O 2)
52
C H A P T E R 2 Forces and Moments In particular, if the density ratio 01/02 is chosen to be equal to the basis, e, of the natural logarithms, one obtains what is called the scale height, h0, of the atmosphere, where (2.51)
ho = R T / g
The change in density with altitude can therefore be expressed roughly by d o / o = - d h / ho
(2.52)
The assumption that the scale height of the atmosphere is a constant in the altitude range of reentry trajectories is only qualitative. Thus, at 120 km altitude (where the temperature is about 360 K, substantially higher than the atmosphere's m i n i m u m temperature of 190 K), the scale height is 9.5 km, compared with only 5.4 km at the temperature m i n i m u m at 90 km altitude, but again 9.2 km at sea level. Nevertheless, because the density changes so drastically from its value of 1.22 kg/m 3 at sea level to 2.22 10 -8 kg/m 3 at 120 km altitude, the concept of a scale height, more or less constant, is a useful one for approximate calculations of reentry trajectories. The equation of motion tangential to the flight path is dV m--d-~- = mgsin y - D
(2.53a)
The corresponding equation for the direction perpendicular to the path is most easily calculated from the path's radius of curvature, rc, and from this the magnitude of the centrifugal force. One shows easily that dy 1 )-1 -T+--COSy
rc--
r
as
where ds is the element of path length. (The first term comes from the change ofthe flight angle, the second term from the change ofthe orientation of the local horizontal.) Balancing the centrifugal force against the gravity c o m p o n e n t perpendicular to the path, and setting ds = V dt, results in the second equation of motion
v
+
cos•
(2.53b)
where the approximation r = rg has been used. It turns out to be advantageous to use the density rather than the time as the independent variable [18]. Making use of (2.52), dt =
ho V sin y 0
so that (2.53a) becomes d(V2/grg) _ 2ho
0
do
-
rg
ohoACD (V2/grg)
msiny
with D = CDA(O/2) V2 (A = cross-sectional area of the vehicle).
53
2.6 Atmospheric Entry
It is convenient here to introduce the constant k = hoACD/ m
(2.54)
and to define in place of the density the nondimensional quantity a = ln(k~)
(2.55)
With this, the final forms of the equations of motion become d(V2/gh~ da (V2/gho)-~a
= 2 - exp(a)(V2/gho ) sin y
(2.56a)
-
(2.56b)
1 - --(V2/gho) rg
cotany
These two equations, with specified initial conditions, suffice to determine the dependent variables V2/gh0 and y as functions of the independent variable a. To solve this system of equations, a numerical method of integration is required. The initial conditions are formulated at a reentry height, h~e, where the effect of the drag on the trajectory starts to become appreciable. For most applications, a convenient choice for entry into the earth's atmosphere is 120 km. Accordingly, the initial conditions are given by (7" --- Ore ----
In (kOsl),
V --
Vre,
]i' -
)/re
(2.57)
Once solved for V and F, the altitude to which they apply can be found by integrating (2.52) between h = 0 and h, with the result that h--
(2.58)
h 0 ( c r s l - - (7)
where r - - " ln(kQsl) and psi the sea level density. Figure 2.14 presents an example, calculated from (2.56), of an earth reentry trajectory. The start of the reentry is taken (arbitrarily) at 120 km altitude, with a velocity of 7.75 km/s and an angle of 10.0 ~ from the local I
I
I
I
I
V = 5.0 ~ ~ . ~ . . . ~
40
3 . 0 J
40 E v
30 ~ 20
+.,
10 4.6
I
I
I
I
4,4 4.2 4.0 3.8 Angle~ (from center of Earth) after entry (deg)
Figure 2.14 Reentry (shown below 50 km altitude) into the earth's atmosphere. Calculated from Eq. (2.56). Entry altitude, 120 km; entry angle y, 10.0~ entry velocity, 7.75 kin/s; k, 100 m 3/kg.
54
C H A P T E R 2 Forcesand Moments
horizontal. The trajectory, which is for k = 100 m 3/kg, is shown as a function of the geocentric angle r (shown in Fig. 2.13), which is a measure of the horizontal distance covered by the vehicle. For a more shallow entry angle, higher entry velocity, a n d / o r with lift, the vehicle can again skip out of the atmosphere. The continuation of the trajectory is then a Kepler-type orbit, which may terminate in another entry, with the vehicle losing velocity each time. The earliest example of a repeated entry-skip-entry maneuver is provided by the NASA Magellan mission to Venus in 1993. The solar panels were turned into the wind to provide aerodynamic braking. After having completed its scientific tasks and entering and skipping about 700 times, with the attitude of the vehicle controlled intermittently by small thruster impulses, the vehicle finally reached the surface and was destroyed (as had been intended).
Nomenclature C
f(v) g
go h
jh k m
r~ n
P r s
A
c, F
G I 11,12,13
Isp In K M N
Pn V
( )ex ( )fm ()g
velocity of light in vacuum Maxwell-Boltzmann velocity distribution gravitational acceleration standard gravitational acceleration altitude scale height solar constant = 1353 W / m 2 Boltzmann constant = 1.3806 10 -23 kg m 2/ (s2 IO; also constant defined in (2.54) mass mass flow rate unit vector, normal to surface pressure position vector; r, radius shear stress area drag coefficient force universal gravitational constant = 6.673 10 -11 m 3/ (kg s 2) incident flux m o m e n t s of inertia; I: inertia tensor specific impulse zonal harmonics; Cnm, Snm"tesseral harmonics Knudsen n u m b e r moment n u m b e r of molecules per unit volume Legendre polynomials; Pnm:Legendre functions velocity nozzle exit free molecule flow gravity; also Earth
Problems
( )h ()i ()r ( )re
( )sl ()t Cr O/h
x tt o ~g
55
solar incident reflected reentry sea level thrust Smoluchowski-Knudsen coefficient; solar radiation absorption coefficient m e a n free path gravitational parameter density angle of incidence gravitational potential
Problems (1) Consider the vertical takeoff of the first stage of a launch vehicle, neglecting the aerodynamic drag, the variation of the gravitational acceleration with height, and the earth's rotation. The liftoff mass is 75,000 kg. The thrust is constant up to an altitude of hi = 25.0 km and is 2.00 times the liftoffweight. The specific impulse is 370 s. Find the time tl after liftoffwhen the vehicle has reached the altitude hi. Also find the vehicle's mass m~ and velocity V~ at that time. (2) A single-stage sounding rocket takes off vertically from sea level. The earth's rotation and the variation of the gravitational acceleration with altitude can be neglected. Aerodynamic drag is included with the simplification that the drag coefficient is assumed to be constant and that the atmospheric density varies exponentially with height, The following data are given: Lift-off mass Propellant mass Thrust Specific impulse Drag coefficient Vehicle cross-sectional area Air density at sea level Atmospheric scale height
200 kg 100 kg 6.00 kN 250 s 0.80 0.0250 m 2
1.225 k g / m 3 9.20 km
(a) By numerical integration of the equation of motion, find the altitude at which the propellant is exhausted. (b) After propellant depletion, the vehicle will coast to a still higher altitude. Find the peak altitude. (3) An asteroid has the approximate shape of an ellipsoid ofrevolution. Its semimajor axis is 6 km, its semiminor axis is 4 kin. The average density is 2.70 103 kg/m 3. Find the surface gravitational accelerations on the semimajor and the semiminor axes. (Universal gravitational constant G = 6.673 10 -11 m 3 kg -1 S-2 .)
56
C H A P T E R 2 Forces a n d M o m e n t s (4) An inflated Mylar sphere of 5.00 m radius is on a low earth orbit at 500 km altitude. The surface density of the skin is 0.70 k g / m 2. (The gravitational parameter of the earth is 3.986 105 km 3/s2; the earth's m e a n radius is 6378 km). Find the gravity gradient torque about the center of the sphere. (5)* A spacecraft in the form of a sphere with 2.00 m diameter is in a low-earthequatorial orbit and is moving from west to east at 300 km altitude. At this altitude, the principal species present are atomic oxygen and molecular nitrogen. Free molecule flow and fully diffuse reflection from the spacecraft surface are assumed. The spacecraft surface temperature is 300 K. The atmosphere corotates with the earth, which has the consequence that the velocity of the spacecraft relative to the atmosphere is 7.30 k m / s (the orbital velocity less the earth's equatorial rotational velocity). Using the following data, c o m p u t e the drag. Number density (m -3) at 300 km Mass (atomic units) Accommodation coefficient
O
N2
7 10~4 16.0 0.80
9 1013 28.0 0.80
A t o m i c m a s s u n i t = 1 . 6 6 0 5 10 -27 kg. Boltzmann
c o n s t a n t = 1 . 3 8 0 6 10 -23 k g m 2 /(s 2 K).
(6)* The problem is related to the production in space of semiconductor and other materials that require for their manufacture an extremely hard vacuum. Advantage is taken of the v a c u u m in the wake of a spacecraft operating in the free molecular flow regime of the earth's atmosphere. Assumed is a spacecraft in the shape of a circular cylinder with radius r - 1.00 m. The axis is parallel to the direction of flight in the atmosphere. The velocity of the spacecraft relative to the atmosphere (i.e., relative to the center of mass velocity of the molecules) is 7.3 km/s. On the axis, at a distance of 0.50r behind the vehicle's base and parallel to it is a small disk with dimensions that are negligible compared with r. The altitude is 400 km. The temperature is 1000 K and the n u m b e r density of the atomic oxygen is 1014 m -3. From the assumed Maxwellian distribution, calculate the n u m b e r of oxygen atoms per second and cm 2 that strike the disk's front surface.
References 1.
Bursa, M. and Pec, K., "Gravity Field and Dynamics of the Earth," Springer Verlag, Berlin, 1993.
2.
Kaula, W., "Theory of Satellite Geodesy," Blaisdell Publishing, Waltham, MA, 1966.
3.
Kellog, O. D., "Foundations of Potential Theory," Springer Verlag, Berlin, 1929.
References 0
0
,
0
0
0
10.
1.
12. 13. 14.
15. 16.
17.
18.
57
Morse, P. M. and Feshbach, H., "Methods of Theoretical Physics," Vol. 2, McGraw-Hill, New York, 1953. Huzel, D. K. and Huang, D. H., eds. "Design of Liquid-Propellant Rocket Engines," Progress in Astronautics and Aeronautics, Seebass, A. R., ed., American Institute of Aeronautics and Astronautics, Washington, DC, 1992. Meyer, R. X., "Inflight Formation of Slag in Spinning Solid Propellant Rocket Motors," Journal of Propulsion and Power, Vol. 8, No. 1, pp. 4550, 1992. Ashley, H. and Landahl, M. T., 'Tkerodynamics of Wings and Bodies," reprint, Dover Publications, New York, 1985. Ashley, H., "Engineering Analysis of Flight Vehicles," Addison-Wesley, Reading, MA, 1974. Martin, E W. and Slotnick, J. P., "Flow Computation for the Space Shuttle in Ascent Mode Using Thin-Layer Navier-Stokes Equations," Progress in Astronautics and Aeronautics, Seebass, A. R., ed., American Institute of Aeronautics and Astronautics, Washington, DC, 1990. Emmons, H. W., ed., "Fundamentals of Gas Dynamics," Section H, Vol. 1, Schaaf, S.A., "Flow of Rarefied Gases," Princeton University Press, Princeton, NJ, 1958. Saksaganskii, G. L., "Molecular Flow in Complex Vacuum Systems," translated from the Russian, Gordon & Breach Science Publishers, New York, 1988. Tribble, A. C., "The Space Environment," Princeton University Press, Princeton, NJ, 1995. Hedin, A. E., "MSIS-86 Thermosphere Model," ]ournal of Geophysics Research, Vol. 92, No. A5, pp. 4649-4662, 1987. Griffin, M. D. and French, J. R., "Space Vehicle Design," AIAA Education Series, American Institute of Aeronautics and Astronautics, Washington, DC, 2nd printing, 1991. Incropera, E P. and Dewitt, D. P., "Fundamentals of Heat and Mass Transfer," John Wiley & Sons, New York, 3rd ed., 1990. Horton, T. E., ed., "Thermophysics of Atmospheric Entry," Progress in Astronautics and Aeronautics, Vol. 82, American Institute of Aeronautics and Astronautics, Washington, DC, 1982. Peake, D. J. and Tobak, M., "Three-Dimensional Interactions and Vortical Flows with Emphasis on High Speeds," North Atlantic Treaty Organization Advisory Group for Aerospace Research and Development, AGARD-AG-252, Paris, 1980. Loh, W. H. T., "Entry Mechanics," Re-Entry and Planetary Entry Physics and Technology, Loh, W. H. T. ed., Springer Verlag, Berlin, 1968.
a This Page Intentionally Left Blank
Orbits and Trajectories in an Inverse Square Field The principal forces that determine the path of a spacecraft are normally gravitation and thrust. In comparison, forces such as atmospheric drag or solar radiation pressure are small and can often be neglected. In this chapter the motion of space vehicles will be considered w h e n gravity is the only force present, that is, at times when the vehicle is coasting without thrust or other forces acting on it. In the vicinity of a planet or other astronomical body, the gravitational field is composed of an inverse square field directed toward the planet's center of mass and a m u c h smaller gravitational perturbation field (Sect. 2.1.3). The latter is caused by the planet's mass distribution, which will generally deviate somewhat from exact spherical symmetry. The resulting perturbation terms, although small, are important in the application of satellites to geodesy and to high-precision navigation. They fall off, however, more rapidly with distance from the center of mass than is the case for the inverse square law force and become negligible at a large distance from the astronomical body. If the astronomical body is a planet, it will orbit the sun and therefore be accelerated relative to a heliocentric ecliptic reference frame (Sect. 1.1.2). A similar case, which is particularly important, occurs for motions of spacecraft in the gravitational and acceleration field jointly produced by the earth and the moon. The present chapter, however, will be limited almost entirely to applications in which the motion of a spacecraft can be described solely in terms of an inverse square law gravitational force produced by a single astronomical body isolated from the influence of other bodies. This approximation is often valid, when the high precision requirements of geodesic or navigational satellites may not be needed. In a final section, some relativistic corrections to Kepler orbits will be discussed. Although the mathematical relations presented in this chapter relate to spacecraft, it should be realized that they are merely a special case of the motion of planets about the sun and of moons about the planets. Nearly all these relations were known to astronomers in the 18th and 19th centuries and even earlier [1, 2]. Modern texts [3-8] on the motion of spacecraft make extensive use of the methods developed originally in astrodynamics. Of course, the mass of a spacecraft can be neglected in comparison with the mass of a planet.
59
60
C HA PT ER 3
3.1
O r b i t s a n d T r a j e c t o r i e s in a n I n v e r s e S q u a r e F i e l d
Kepler Orbits and Trajectories Newton's second law therefore applies in the form d2r dt 2
/z = ---r
(3.1)
r3
where r is the position vector from the center of attraction to the spacecraft a n d / z the gravitational p a r a m e t e r of the astronomical body. Since the position vector is in the plane defined by the center of attraction and the acceleration vector, it is evident that the m o t i o n is in a plane t h r o u g h the center of attraction. Two constants of m o t i o n are the orbital a n g u l a r m o m e n t u m , designated by h, of the spacecraft a b o u t the center of attraction, and the sum, designated by w, of the potential and orbital kinetic energies. (A third constant of m o t i o n is the eccentricity vector, perpendicular to the orbit plane and of m a g n i t u d e e, as defined below later.) Both h and w are u n d e r s t o o d to be taken per unit mass of the spacecraft. Therefore, if v denotes the velocity of the center of mass of the spacecraft in an (approximately inertial) reference frame in which the planet is at rest, h = r x v w -
(3.2) + ~11)2
-u/r
(3.3)
If, as indicated in Fig. 3.1, 0 is the angle b e t w e e n r and a direction fixed in the reference frame, the radial and tangential c o m p o n e n t s of the velocity are 1)r -'- r ,
1)o =
rO
so that h = r20 - const.
(3.4)
and w -
-lz/r
+ ~1 (1:2 +
r202)
-
const.
F
:is ath
Figure 3.1 Spacecraft path in an inverse square central force field.
(3.5)
3.1 Kepler Orbits and Trajectories
61
Substitution of O from (3.4) into (3.5) and differentiation with respect to time result in the relation
J~4- Ix/r 2 -- h2/r 3 - - 0 It can be cast into a more standard form by introducing the reciprocal radius u = r - ~. Since
dr dO dr h _ h dU f = dO dt = dO r 2 = dO .~
r--
d?dO_
d(hdU)
do a t -
h
_h2u2 d2u
-F=
u is seen to satisfy the h a r m o n i c oscillator equation d2 u dO 2
+ u =
ix -
(3.6)
-
h2
with the solution u = / z / h 2 + C cos(O - Oo)
(3.7)
where C and 00 are constants of integration. The constant C can be expressed in terms of the constants of motion h and w by observing that from (3.4), (3.5), and the expression just found for u
12[(u)2 ]
W --" --flU-~- ~
~-~
-'1- U2
= ~1( C 2 h2 _ / . / , 2 / h 2)
hence
tx ~ 2wh 2 C=--l--~--5 1 + ~2
h# 0
(3.8)
For physically realizable systems, u, and therefore C, m u s t be real. Hence 2wh2/# 2 > -1
(3.9)
For positive C, u has a maximum, hence r a m i n i m u m at 0 - 00. For negative C, the m i n i m u m r occurs w h e n 0 = 00 + rr. This point on the trajectory where the spacecraft is closest to the astronomical body is, in conformance with astronomical custom, referred to as the periapsis (in the case of the earth usually referred to as perigee, in the case of the sun as perihelion). For closed orbits, the point of largest distance is called the apoapsis (or apogee in the case of the earth and a p h e l i o n in the case of the sun) (Greek: p e r i . . . = around, about; a p o . . . = out of, from). It is customary to count the angle 0 in the direction of motion starting from the periapsis. Therefore 00 = 0 for positive C, 00 = -Jr for negative C. For a circular orbit, the origin of 0 is usually taken at its intersection with either the equatorial or the ecliptic plane. In astronomy, 0 is known as the t r u e a n o m a l y and is often given the symbol f in place of 0.
62
C H A P T E R 3 Orbits and Trajectories in an Inverse Square Field
EllipticOrbit-~
A
F
P
Hyperbolic Trajectory
"%~\ ~/~\
x
F
|
//[ \\ /a -----~]
x F"
/
>7
o<]
Figure 3.2 Geometry of the one-body problem. C, center of attraction; S, spacecraft; P, periapsis; A, apoapsis; E F', focal points. From (3.7) follows the final result for describing the path, r =
h2
/z(1 + e c o s O ) '
h # 0
(3.10)
e > 0
(3.11)
(irrespective of the sign of C), where e = v/] + 2 w h 2 / # 2,
Equation (3.10) is the expression in polar coordinates of a conic section, with e the eccentricity. For 0 < e < 1, the path is an ellipse, for e > 1 a hyperbola, and in the two limiting cases a circular orbit and a parabolic path. It also follows that the center of the astronomical body is a focal point of the conic section. Figure 3.2 illustrates these geometrical properties. It is of historical interest to note that expressions equivalent to (3.10) were derived by Johannes Kepler (1571-1630) on the basis of Tycho Brahe's (1546-1601) astronomical observations of the motion of the planets. Kepler formulated these laws: (1) The orbit of each planet is an ellipse in a plane containing the sun, with the sun at one focus. (2) The line connecting the sun and the planet sweeps out equal areas in equal time intervals. (3) The square of the orbital period of a planet is proportional to the cube of the semimajor axis of the ellipse. Newton was the first to derive Kepler's three laws mathematically, starting from his law ofuniversal gravitation. Although he had at that time already invented the calculus, he used instead geometrical arguments. Kepler's laws are actually more accurate w h e n applied to spacecraft than to planets. This is because planetary masses are not entirely negligible when compared with the sun's mass. (For Jupiter, the most massive of the planets,
3.1 Kepler Orbits and Trajectories
63
the mass ratio is approximately 0.955 10-3.) As a consequence, the foci of the planetary orbits do not exactly coincide with the center of the sun, as had been assumed by Kepler, but more nearly with the center of mass of the solar system, the so-called barycenter. Returning to the earlier development, one finds a remarkably simple relation between the semimajor axis, denoted by a, and the energy w. At the periapsis the velocity is purely transverse and 0 = 0, so that from (3.3) and (3.10) /z w=---+ rp
1 h2 2 r~'
rp=
h2
/z(1 + e)
where rp is the periapsis radius. Hence w = -/x(1 - e)/(2 rp). From the geometrical relation indicated in Fig. 3.2, rp = +a(1 - e), it follows that w = ~tx/(2a)
(3.12)
h 2 = +/za(1 - e 2)
(3.13)
and also
In these equations and in all others in this chapter where two signs are indicated, the upper sign stands for elliptic orbits, the lower one for hyperbolic trajectories. (The reader needs to be aware that with some authors the quantity a is negative for hyperbolas. In the present text, the convention is adopted that a is always positive.) From (3.4) and (:3.10) follows Vr = (/Z / h) e
sin O,
vo = (/z / h) (1 + e c o s 0)
~
(2
1)
h#0
(3.14)
r :a If dA designates the area enclosed between two radii separated by the angle dO, then from (3.4), d A / d t = hi2. Hence f l t~ dA = ~h (tz - tl)
(3.15)
which states that for a given orbit the radius vector r sweeps out equal areas in equal time intervals. This equation is known as Kepler's "second law." (Actually, it was discovered by Kepler before the "first law.") An important conclusion drawn from it is that satellites on eccentric orbits move more slowly near their apoapsis than they do near their periapsis. Broadcasting systems, such as the Molniya (Russian: lightning) system, which are designed to cover primarily northern latitudes, take advantage of this by using a large eccentricity and by placing the apogee north of the equatorial plane. As a consequence, a typical ground receiver can have a direct line of sight to the spacecraft for a time that can exceed by more than one-half the orbital period. Kepler's "third law" applies to closed orbits, that is, to elliptic and circular paths. Let P designate the orbital period. Because the area of an ellipse with
64
C H A P T E R 3 Orbits and Trajectories in an Inverse Square Field
s e m i m a j o r axis a a n d s e m i m i n o r axis b = a~/1 (3.15) t h a t
e 2 is ~rab, it follows from
P = 2Jr(a2/h)v/1 - e 2 F r o m (3.10) a n d the g e o m e t r i c definition of the eccentricity (Fig. 3.2), rp = h2/[~(1 + e2)] = a(1 - e)
so that
P = 2~ v/a 3//~
(3.16)
proving Kepler's third law. Remarkably, the p e r i o d d e p e n d s only on the semim a j o r axis a n d n o t on the eccentricity. Alternatively, the p e r i o d can be expressed in t e r m s of the total energy, w, per unit m a s s of the spacecraft, b e c a u s e from (3.12)
P - ~/~r189 (-w) -3
(3.17)
again showing the result to be independent of the eccentricity. Figure 3.3 illustrates the relationship between the total energy and the orbital periods of several earth satellites. (Two of the examples shown have circular orbits; the Molniya satellites have strongly eccentric orbits; the moon's orbit also has an appreciable eccentricity: its distance from the earth's center varies from 356,400 to 406,700 km.) Finally, we note that for parabolic paths, e = 1 and w = 0. Because at an infinite distance the potential energy tends to zero, it follows that the kinetic energy also tends to zero. The parabolic trajectory therefore represents the limiting case in which a spacecraft will escape the gravitational field, with a velocity that tends to zero at infinity.
-0.5
--
n
--- -1 E
v O
-2
// in ior t
oO t~ t~q
/ G e o s l a t i o n a r y Orbit
-5
t-C--
'" -10 -20
/~300 1
"1
2
,
km Altitude Orbits ,
I,,,,1
5
10
I
i
~ I t~lJl
I
20 50 100 200 Orbit Period(h)
,
,
I ,,,,1
500 1000
Figure 3.3 Total energy per unit mass of satellite versus orbital period.
3.2 Position as a Function of Time
65
The m i n i m u m velocity at periapsis that is n e e d e d for escape, if r0 is the distance of the periapsis from the center of the planet, is, from (3.3), Vesc = ~/2/z/ro
(3.18)
This can be c o m p a r e d with the velocity Vcrc = ~/'/Z/t'0 of a spacecraft on a circular orbit with the same radius r0. Hence Vesc = ~/2 Vcrc. For the earth, a s s u m i n g an initial circular orbit with r0 = 6678 k m (the m e a n equatorial radius plus 300 k m altitude, a value that is fairly typical for parking orbits), Vcrc is approximately 7.73 km/s, the escape velocity 10.9 k m / s . For a westto-east launch at the equator, the earth's circumferential velocity due to its diurnal rotation is 0.465 km/s, so that the m i n i m u m theoretical velocity i n c r e m e n t that is n e e d e d for escape is 10.4 km/s.
3.2
Position as a Function of Time So far, only the path, n o t the position of the spacecraft as a function of time, has b e e n considered. Whereas the p a t h can be expressed in terms of elementary functions, this is n o t the case for the t e m p o r a l d e p e n d e n c e of position. We limit the discussion at first to elliptic orbits. It is c u s t o m a r y to introduce the so-called eccentric a n o m a l y , an angle that is traditionally designated by E. As shown in Fig. 3.2, the eccentric a n o m a l y is o b t a i n e d by projecting the satellite position on a circle of radius a that passes t h r o u g h the periapsis and apoapsis. The n e e d e d relation is m o s t easily d e m o n s t r a t e d by introducing, as indicated in Fig. 3.2, the rectangular Cartesian coordinates of the satellite position x = a(cos E - e),
y = a~/1 - e 2 sin E
(e < 1)
(3.19)
from which r = a(1 - e cos E) a n d a(cos E
e) = r cos 0,
av/1 - e 2 sin E = r sin 0
(e < 1)
(3.20)
(When using a single one of these equations to d e t e r m i n e 0 from E, or E from 0, an ambiguity arises as a c o n s e q u e n c e of the multivaluedness of the inverse trigonometric functions. In simple cases, the correct q u a d r a n t of the c o m p u t e d angle is often obvious. But w h e n it is e m b e d d e d in software, it is advisable to resolve the ambiguity by using b o t h equations. This procedure also provides an easy check on the calculation.) The angular m o m e n t u m , h, per unit mass, therefore b e c o m e s h = x y - y:~ -- a2v/1 - e 2/~(1 - ecos E) - av/1 - e 2 r/~ C o m p a r i s o n with (3.13) results in/~(1 - ecos E) = V/t~/a 3, or integrated E - esin E - v / t t / a 3 ( t - tp)
(e < 1)
(3.21)
where tp is the time of periapsis passage. This e q u a t i o n relates satellite position and time a n d is k n o w n as Kepler's e q u a t i o n . In a s t r o n o m y the (dimensionless) r i g h t - h a n d side of Kepler's e q u a t i o n is traditionally designated by M a n d is called the m e a n a n o m a l y . The factor
66
C H A P T E R 3 Orbits and Trajectories in an Inverse Square Field v/lz/a 3 is designated by n and is called the m e a n angular velocity [since, as follows immediately from (3.16), n is the angular velocity averaged over the elliptic orbit]. Therefore (3.21) is often written in one of the forms
E - e s i n E = n ( t - tp) = M
(e < 1)
Before discussing methods for solving Kepler's equation, we complete this section by listing some other frequently used equations. Thus it follows from differentiating Kepler's equation, solving the result for/~, and substituting the result into (3.19) after differentiation that !
sin E 1 --- ecos E'
x= -
/ / z ( 1 - e 2) cos E a 1 - e cos E
)
V
(e < 1)
(3.22)
Analogous results can also be obtained for the hyperbolic trajectories. The eccentric anomaly, in this case called the hyperbolic anomaly, can be defined by m e a n s of the relation r = - a ( 1 - e cosh E) analogous to the equation preceding (3.20). We list these results: x = a(e - cosh E),
y = ax/e z - 1 sinh E
E - esinh E = - x / t t l a a ( t - tp)
(e > 1) (e > 1)
(3.23) (3.24)
[
= _~_
Va
3.3
sinh E e cosh E - 1'
)
/ / z ( e 2 - 1) cosh E a e cosh E - 1
V
( e > 1)
(3.25)
D'Alembert and Fourier-Bessel Series The need frequently arises to compute a spacecraft's position in a known orbit as a function of time. This requires the solution of Kepler's equation (3.21) or (3.24), hence of a transcendental equation for the eccentric anomaly. (The inverse problem of finding the t i m e at which the spacecraft is in a given position is straightforward because Kepler's equation is linear in time.) Numerically, the solution can be found by the Newton-Raphson or the more recent, robust m e t h o d of Muller. Once the eccentric anomaly is found, the position follows directly, for instance, from (3.19) or (3.23). Alternatively, the position of the spacecraft can be found by series expansions, originally developed for applications in celestial mechanics.
3.3.1
D'Alembert's Method This series expansion (D'Alembert, 1717-1783) is useful for elliptic orbits with eccentricities not too close to unity. In astronomy, particularly because the orbits of the majority of planets and of their moons have only small eccentricities, D'Alembert's m e t h o d has found very frequent applications. It is also useful in applications to space technology.
3.3 D ' A l e m b e r t a n d Fourier-Bessel Series
67
Let e = E - M. Clearly, e is a periodic, odd function of t (with t = 0 at a periapsis passage), hence of M. Making use of e, Kepler's equation can be written e = E - M = e s i n E = e s i n ( M + e) = e ( s i n M c o s e
+
cos M s i n e )
Since e ~ 0 as e ~ O, this suggests expanding e in a power series in e, e = Oll ( M ) e -47ol2(M)e 2 + 0r (M)e 3 + . . . with coefficients Cti that are 0(1) and dep e n d periodically on M. Substituting the series into both sides of the preceding equation,
o~le + c~2e2 + . . . = e[sin Mcos(c~le + c~2e2 + . . . ) + cos
Msin(otle +
o~2e2 +...)]
The series arguments in the sines and cosines are close to 0 for sufficiently small e, which allows the sines and cosines to be expanded with rapid convergence by their own power series. Collecting the terms with the same power in e, one finds al = sin M, c~2 = al cos M ~1 sin(2M)
0r3 = --~1 sin M + 3 sin 3M,
.....
The derivation is completed by substituting the coefficients ai back into the power series for e and collecting terms with the same frequency in M. The more complete result, carried out to higher orders in e than sketched here, is
E = M + e ( 1 - ~ le2 + A
19
e4 ~ s i n M + e ]
2( 1 le2 ) s i n 2 M 2-6
2 7 )e 2 sin 3 M + le4 sin 4 M + 125e5 sin 5M + ... 128 5
+e3(3 8
(3.26) The functions 0 - M and r are also periodic functions of M, the first one odd, the second one even. We only state the results. Neglecting, as before, terms of order e 6 and higher, 0 = M( + e
2
12
41 e 2 + 5 ) e 4 s i n M + e2 ( 5 - l2l4e 2 ~,) s i n 2 M
10 e4 s i n 4 M
64 e 2 s i n 3 M + - ~
1097
+ 960 ea sin 5M + . . .
(3.27) and
r a-
--
1 +
le2 2
_e3(3
-- e
(1
--
3e2 + 5 8
192
e 4~
]
cos M
--
e 2 ( ) c-o s 12 eM2 1
45) i-2-8e2 cos 3 M - 3 e4 cos 4 M -
2
125 5 3--~e cos 5M . . . .
(3.28) It has been shown [9] that the three series (3.26), (3.27), and (3.28) converge for all values of M if e < 0.6627.
C H APT E R 3 Orbits and Trajectories in an Inverse Square Field
68
3.3.2
Fourier-Bessel Series In addition to D'Alembert's series, a n u m b e r of other methods, suitable for different ranges of the eccentricity, exist for solving Kepler's equation. Sketched here briefly is the Fourier-Bessel series. It converges rapidly for small values of e. By differentiating Kepler's equation (3.21) for elliptic orbits, one obtains
dE =
dM 1 - e cos E
Since (1 - ecos E) -1 is a periodic, even function of M w i t h period 2n, it can be expressed by the Fourier series oo
1
1 - e cos E
= Ao + Z
Ak cos(kM)
(3.29)
k=l
where
1 f+~ A0-
~
dM
1 f+~ dE -- 1
~ 1-ecosE
2yr J_rr
1 f + ~ cos(kM) d M = -1 f + ~ cos [k(E - e sin E) ] dE Ak = -1 ecosE 7r Jr Yr 7r = 2]k(ke)
k - 1, 2 . . . .
and where ]k is the Bessel function of the first kind of order k. The final result is obtained by integrating (3.29) over M term by term, resulting in
E = M + 2Z
~]k(ke)sin(kM)
(3.30)
k=l
3.4
Orbital Elements Up to this point it was sufficient to specify elliptic orbits or hyperbolic trajectories by their semimajor axis and eccentricity. However, for a complete description of the path of a spacecraft, relative to a fixed center of attraction, additional data are needed: Two angles are required to specify the orientation of the plane of the motion, and one angle is needed to specify the orientation in this plane of the major axis. Finally, the time at which the spacecraft passes a fixed point needs to be specified. The six scalar quantities referred to are known as the o r b i t a l e l e m e n t s (also known as the "Keplerian elements"). Those most often chosen in space technology are: a: the s e m i m a j o r axis of elliptic orbits or hyperbolic trajectories. e: the e c c e n t r i c i t y . i: the inclination, that is, the angle between the plane of the motion and the equatorial plane of the earth or planet. For motions about the
3.4 Orbital Elements
69
sun, in place of the equatorial plane the ecliptic plane is used. A more precise definition, which avoids a possible ambiguity in the sign of i, is to define the inclination as the angle i, 0 _< i < n, between the angular m o m e n t u m vector (therefore a vector perpendicular to the orbit plane) and the northward-pointing perpendicular to the equatorial or ecliptic plane. ~" the right ascension of the ascending node, which is the angle from the vernal equinox line to the ascending node, taken positive in the easterly direction. w: the angle from the ascending node to the periapsis, taken positive in the direction of motion. tp" the time at which the spacecraft passes the periapsis. (Sometimes tp is replaced as an orbital element by the m e a n anomaly M at some reference time.) Figure 3.4 illustrates the orbital elements. Point P in the figure is the
periapsis, point AN the ascending node, that is, the point where the spacecraft crosses the equatorial or ecliptic plane from the southern to the northern hemisphere. The angles fZ, i, w (in this order) correspond to the Euler angles r 0, ~ shown in Fig. 1.7. For historical reasons, and in conformity with the technical literature, they are denoted by these new symbols. Some of the orbital elements, as listed, can be ill defined. Thus, for circular orbits, for which there is no distinct periapsis, and also in cases in which e ~ 0, it is customaryto measure the position ofthe spacecraft instead from the ascending node. If i = 0, there is no distinct ascending node. In
q.
Z~
h~
AN
a j1
qf
- e 2
Equatorial or Ecliptic Plane Orbit
Figure 3.4 The six orbital elements a, e, i, fl, o~, tp. ~ , first point of Aries; N, north; AN, ascending node; P, periapsis; h, angular momentum vector.
70
C H A P T E R 3 Orbits a n d Trajectories in a n Inverse Square Field this case, or w h e n i ~ 0, o n e often uses t h e q u a n t i t y & = ~o + fl as o n e of the orbital e l e m e n t s . In t h e i m p o r t a n t case of circular orbits in the e q u a t o r i a l plane, e = i = 0. In this case j u s t t w o orbital e l e m e n t s suffice: the orbit r a d i u s a a n d the t i m e of p a s s a g e of the v e r n a l e q u i n o x line. In t h e inverse s q u a r e force field c o n s i d e r e d in the p r e s e n t chapter, t h e orbital e l e m e n t s are c o n s t a n t s . Their i m p o r t a n c e , however, derives largely f r o m t h e i r u s e in a n a l y z i n g t h e effects of small p e r t u r b a t i o n s f r o m the inverse s q u a r e field. Such p e r t u r b a t i o n s m a y be caused, for instance, by the n o n s p h e r i c i t y of t h e m a s s d i s t r i b u t i o n of the earth, as d i s c u s s e d in Sect. 2.1.3. In t h e s e cases, t h e orbital e l e m e n t s are n o l o n g e r c o n s t a n t b u t c h a n g e slowly w i t h time, p r o d u c i n g s e c u l a r t e r m s in the e q u a t i o n s . Finally, t h e following e x p r e s s i o n s for t h e d e c l i n a t i o n 8 a n d the right a s c e n s i o n ~ (Figs. 1.3 a n d 1.4) in t e r m s of the orbital e l e m e n t s i, f2, co a n d the t r u e a n o m a l y 0 m a y b e n o t e d :
sin 8 = sin i sin(w + 0) cos(a - fl) =
cos(oJ + O) cos6
(3.31a) (3.31 b)
T h e y follow directly f r o m c o n s i d e r i n g t h e right angle s p h e r i c a l triangle b o u n d e d by the orbit, t h e equator, a n d t h e m e r i d i a n t h r o u g h the spacecraft position. T h e s a m e e q u a t i o n s also a p p l y w h e n the s y m b o l s are i n t e r p r e t e d to a p p l y to the ecliptic as the reference plane. F r e q u e n t l y u s e d c o o r d i n a t e t r a n s f o r m a t i o n s b e t w e e n elliptic orbital ele m e n t s a n d o r t h o n o r m a l c o o r d i n a t e s are listed in Tables 3.1 a n d 3.2.
Table 3.1 Transformations from elliptic orbital elements to orthonormal coordinates and their derivatives (For the definition of the coordinates, see Fig. 3.4) Defining
x = :~ = p =
=
11 = cos co cos fl - sin co sin f2 cos i lz = - s i n co cos f2 - cos co sin fl cos i /3 - sin f2 sin i ml = cos co sin ~2 + sin co cos ~2 cos i mz = - s i n co sin f2 + cos co cos f2 cos i m3 = - c o s fl sin i nx = sin co sin i nz = cos co sin i /'/3 = cos i z = nl~ +n2~ 11~ + 12r#, y = mx~ + m20, (Iz/h)[-lx sinO+ 12(e + cosO)] ( ~ / h ) [ - m a sinO+ m2(e + cosO)] (/z/h)[-nx sinO+ n2(e + cosO)]
3.5 Spacecraft Visibility above the H o r i z o n
71
Table 3.2 Transformations from orthonormal coordinates and their derivatives to elliptic orbital elements (For the definition of the coordinates, see Fig. 3.4) Defining r = v/x2 + y2 q_ Z2 r(g 2 + ~2 + 22))
a =/zr/(2/z-
e 2 = (1 - r/a) 2 + (/za)-z (xg + y~ + z2) a h = v//za(1 - e ~-) cos i = ( x ~ - y ~ ) / h sin f~ = (y2
(0 <_ i < Jr)
zyz)/(h sin i),
esinE = ( / z a ) - Z / 2 ( x ~ + y ~ + z ; ~ ) , O = 2 t a n -~
(jx+e 1
cos f~ = (x2 - z 2 ) / ( h s i n i) ecosE = 1 - r / a
etan
z sin(o) + O) - r sin ., z
cos(o)+ O) = (y/r) sin f2 + (x/r) cos f2 i
3.5
Spacecraft Visibility above the Horizon As an example of the use of orbital elements, we calculate the time at which a spacecraft b e c o m e s visible above the horizon of a g r o u n d - b a s e d observer or g r o u n d station. Problems of this type are c o m m o n l y encountered, for instance, in designing a satellite navigational system in which several satellites (typically four) m u s t be simultaneously viSible to the observer. The same p r o b l e m also arises in satellite c o m m u n i c a t i o n systems, because at the microwave frequencies used in these systems, the c o m m u n i c a t i o n paths are essentially straight line. In the following we calculate the limit of visibility as d e t e r m i n e d by the theoretical horizon, that is, by the plane t a n g e n t to the spherical figure of the earth at the location of the observer. In practice, a t m o s p h e r i c a b s o r p t i o n a n d refraction at the carrier frequency, or simply obstacles such as m o u n t a i n ranges or buildings, limit the useful paths for c o m m u n i c a t i o n s to s o m e positive angle of elevation above the horizon. Unless special circumstances dictate otherwise, 10 ~ to 15 ~ above the horizon are considered to be a m i n i m u m for m o s t microwave links. Figure 3.5 indicates the geometrical relations. A spherical earth is assumed. The g r o u n d station, designated by O, is located at geographic latitude q~0 and longitude )~0 east of Greenwich. The orbital plane intersects the horizon plane of the g r o u n d station along the line H-H'. The spacecraft b e c o m e s visible at H and disappears at H'. Orthogonal coordinates are defined, with the origin at the earth's center, the x axis coincident with the vernal equinox line, a n d the z axis along the n o r t h w a r d polar axis. To find the direction of the line H-H', we introduce the unit vectors no perpendicular to the g r o u n d station's horizon plane, pointing outward from the earth, a n d ns p e r p e n d i c u l a r to the plane of the spacecraft orbit, pointing in the direction of the spacecraft's angular m o m e n t u m vector. The vector
72
C H A P T E R 3 Orbits and Trajectories in an Inverse Square Field
nO /[
7 -% Earth-x/ /
D
Equatorial Plane
Figure 3.5 Visibility of a spacecraft for a ground observer at O. Spacecraft is visible above line H-H'. NH = no x ns t h e n d e t e r m i n e s the line of i n t e r s e c t i o n of the t w o p l a n e s a n d p o i n t s t o w a r d p o i n t H. Ifa0 d e s i g n a t e s the right a s c e n s i o n of the g r o u n d station's m e r i d i a n a n d a c the right a s c e n s i o n of G r e e n w i c h , t h e n ao = aG + Xo. The c o m p o n e n t s of no are
nox = cos q~ocos ~0,
n0y = cos ~0 sin a0,
no~ = sin ~0
M a k i n g use of t h e orbital e l e m e n t s i n t r o d u c e d in the p r e c e d i n g section, we o b t a i n for the c o m p o n e n t s of ns
nsx = sin i sin ft,
nsy = - sin i cos S2,
nsz = cos i
Hence NHx = COS i cos q~0 sin ct0 + sin i cos S2 sin q~0 NHy = sin i sin f2 sin q~0 - cos i c o s r
cosa0
(3.32)
NH~ = - s i n i cos q~o cos(f2 - ~0) The u n i t v e c t o r NA p o i n t i n g t o w a r d the a s c e n d i n g n o d e A h a s the c o m ponents NAx = cos f2,
NAy = sin S2,
NA~ = 0
(3.33)
The angle ~o + OH in the p l a n e of t h e orbit, w h e r e OH is the t r u e a n o m a l y of p o i n t H, is t h e n o b t a i n e d f r o m the scalar p r o d u c t (NH/NH) 9NA. H e n c e
(NHx/NH) COS S2 + (NHy/NH) sin ~2 = cos(~o + 0H)
(3.34)
If the orbital e l e m e n t s a n d the l o c a t i o n of t h e o b s e r v e r are k n o w n , this
3.6 Satellite Observations and the f and g Series
73
then determines OH. From (3.10) and (3.20) follow the radius rH and the eccentric anomaly EH at point H, and from (3.21) the time tH at which the spacecraft rises above the horizon. As is seen from this derivation, all six orbital elements need to be known when calculating tH. The same method also provides the corresponding results for the descent of the spacecraft below the horizon.
3.6
Satellite Observations and the f and g Series A frequently occurring task is to observe from a ground station the position and velocity of a spacecraft and from these to derive the orbital elements, for instance, after injecting a spacecraft into a low-altitude parking orbit. Injection errors, usually very small but inevitable, can then be corrected by programming the thrust direction and (for liquid propellant engines) the duration of the burn of the next rocket stage. Alternatively, the correction can be made by the spacecraft's own thrusters. Precision observations of spacecraft orbits use a combination of optical and radar techniques. More accurate than any other optical technique is the technique of photographing the spacecraft against a star background. The photographs are repeated at precise instants oftime, obtained from standard time signals. For high precision, a number of corrections may have to be made, which may include changes in the right ascension and declination of the stars caused by the earth's precession and nutation. These may have accumulated since the time of compilation of the star catalog that is being used. Other corrections may include the effect of differential atmospheric refraction between a star position and the spacecraft image, an effect that can become significant when the spacecraft is near the horizon. The use of theodolites provides an alternative, although slightly less accurate, method. All optical techniques provide highly precise measurements of the true azimuth and elevation of the spacecraft at fixed times and of the velocity component perpendicular to the line of sight. Because of the limited angular resolution of the beam, radar measurements cannot give information at right angles to the line of sight with accuracies comparable to optical methods. However, radar complements optical determinations by providing highly accurate range and range rate measurements. The range, that is, the distance along the line of sight, is obtained from the time delay between pulses transmitted by the ground station and the return pulses received after reflection from the spacecraft ("skin tracking"). The radar pulses reaching the spacecraft can also be amplified by a repeater on the spacecraft and returned to the ground station. Refraction in the troposphere and ionosphere causes the propagation velocity to differ slightly from the speed of light in vacuum. This effect depends on the angle of elevation and on tropospheric and ionospheric conditions but can be corrected by suitable calibration or by the use of two different, widely spaced, microwave frequencies. The range rate, which gives the spacecraft's velocity component relative to the ground station along the line of sight, is obtained by a determination
74
C H A P T E R 3 Orbits a n d Trajectories in an Inverse Square Field ofthe Doppler shift between the transmitted and returned signals. The measurement consists of a count of the cycles over a fixed time interval and can be made extremely accurately. By using more than one ground station so as to obtain at the same instant several range and range rate measurements, radar can replace entirely, although with somewhat less accuracy, optical measurements. The so-called f and g series that are discussed in the following are important in several methods of orbit determination. They are also useful in extrapolating over short time intervals a known spacecraft position and velocity, a task that occurs in planning a spacecraft rendezvous. The results apply to elliptic orbits as well as to hyperbolic trajectories. In deriving the method, it is assumed that the position vector r0 and velocity vector v0 are known at some initial time to. The method then allows one to find r and v after a time interval At later. Expanding r in a Taylor series in At,
r=ro+
(*) ~
) (at) o A t + ~1\ d(a2r t 2 o
+
1(d3r)
-dF
o
(at)3
1 (d4r) (At) 4 -[-... + ~ k , dr4
(3.35)
o
we calculate the coefficients up to order four. Expansions to higher order can be found in more specialized texts, for example, in Herrick [10]. Because a Keplerian orbit or trajectory is assumed, d 2 r / d t a can be replaced by the right-hand side of (3.1). Higher derivatives then follow by differentiation; thus dar dt 3 =
3#f /z r4 r-~--~v
d4r
1#2( /( z? ) 2 -
dt 4
r4
r2
r
F3~) r + -6/zf r4v
When substituted into the series, the as yet unknown term (Do will appear. By differentiating ? - (r/r) 9v one finds _ _r-l(lz/r
+ (~)2 _ v 2)
It will be found convenient to separate in the series terms that are proportional to r0 from those proportional to v0 and to write r-
fro + gvo
(3.36)
Upon substitution of these results into (3.35), one finds the final result /z
g -
At-
/~
Uro (At3
/zfo
+
/~fo
3
24ro~ 70 + 15(r~
(At34 . . . .
)
- 3v2 (At)4 + " "
(3.37)
3.7 Special Orbits
75
Similar expressions can also be o b t a i n e d for the velocity, because on differentiating (3.36) one finds v = ( d f / d t ) r o + (dg/dt)Vo
(3.38)
where
al/at = -Fo At +
3j/,r0
2
(At) _
/zr0
/z (At) 2 + ~ ( A t ) ag/at= l - 5-~r~ r~
~
3
+ 15(fo) 2 -- 3Vo 2 (At) 3 +...
(3.39)
The following n u m e r i c a l example illustrates the accuracy of these approximations. Assuming an elliptic orbit a b o u t the earth, of eccentricity e = 0.5 and altitude at perigee 300 k m above the m e a n equatorial earth radius, the period P = 15,337 s. We furthermore a s s u m e that the initial position of the spacecraft was at the true a n o m a l y 00 = 90 ~ (at this point, as is easily shown, the angle b e t w e e n the t a n g e n t to the orbit a n d the radial direction is just cotan -1 e). If we a s s u m e At -- P/100, the ratio of the last t e r m listed in (3.37) to the first term is - 7 . 3 0 10 - 6 for the f series a n d 1.13 10 - 4 for the g series. If A t = P/IO, the c o r r e s p o n d i n g values are - 0 . 0 7 3 a n d 0.113, respectively. It also follows from (3.36) that for At = P / 1 0 0 the last t e r m listed in the f series c o n t r i b u t e s to the final position a distance in the initial radial direction o f - 7 3 m, b u t - 7 3 0 k m if At = P/IO. The last t e r m listed for the g series contributes c o r r e s p o n d i n g values of 122 rn if At = P / 1 0 0 b u t 1225 k m if At = P/lO. For ease of reference, the m o r e i m p o r t a n t equations derived up to this point in this chapter are s u m m a r i z e d in Table 3.3.
3.7
Special Orbits In this section s o m e special Kepler orbits that have found frequent applications are considered.
3.7.1
Circular Orbits From (3.14), with a = r, follows the velocity, Vcrc, of a spacecraft on a circular orbit of radius r Vcrc -- ~
(3.40)
For instance, for a low e a r t h orbit, a s s u m i n g an altitude of 300 k m above the earth's m e a n equatorial radius (6378.1 km), yore = 7.726 k m / s . The siderial period, from (3.16), is 5431 s or 90.5 min. To find, however, the time interval b e t w e e n two successive crossings of the same meridian, the rotation of the earth m u s t be considered (see Fig. 1.9). Assuming that the orbit is in the
76
C HA PT ER 3
Orbits a n d Trajectories in a n I n v e r s e S q u a r e Field
Table 3.3 Kepler trajectories, s u m m a r y of formulas (For the definition of the symbols, see Chap. 3 Nomenclature) Ellipse
Hyperbola
h2 r=
Same
~(1 + e cos O)
e = v/1 + 2 w h 2 / i x 2
Same
w = -/z/C2a)
w = +/z/(2a)
h 2 = +/za(1 - e 2)
h 2 = -/za(1 - e 2)
v = v//z(2/r-
v = v/lZ(2/r + i / a )
i/a)
P = 2n v/a 3/lz
E - esin E = ~
(t - tp)
E - esinh E = - ~
x - a(cos E - e)
x = a(e-
y = a~/1 - e 2 sin E
y = a~/e 2 - 1 sinh E
r = a(1 - e cos E)
r = - a ( 1 - e cosh E)
= - ~
sin E/(1 - e cos E)
= - ~
y = ~ - a ~ / 1 - e 2 cos E / ( 1 - ecos E) # (At) 2 + /zf0 f = 1 - ~r~ 2-~4 (At)3 x
g~--
~ + 15(?o) 2 ro
#
A t - - ~T03 (At)3
-~-
- 3v #fo
( t - tp)
cosh E)
sinh E / ( e c o s h E - 1)
fl = ~ / - ~ - ~ / e 2 -
1 cosh E/(ecoshE-
1)
/z 24ro5 (At) 4 - I - " "
-~-~ (At) 4
....
Same Same
e q u a t o r i a l p l a n e a n d f r o m w e s t to east, this t i m e interval in this e x a m p l e b e c o m e s 90.5 + 6.1 - 96.6 m i n . T h e velocity is s e e n to d e c r e a s e w i t h i n c r e a s i n g radius. T h u s t h e velocity of the m o o n o n its orbit ( w h i c h is at least v e r y r o u g h l y circular) a b o u t the e a r t h is o n a v e r a g e only a b o u t 1.0 k m / s . It also follows f r o m (3.40) t h a t n o t only t h e o r b i t i n g velocity b u t also the c o r r e s p o n d i n g a n g u l a r velocity d e c r e a s e s w i t h i n c r e a s i n g radius. Hence, s p a c e c r a f t o r b i t i n g at a larger d i s t a n c e f r o m t h e c e n t e r of a t t r a c t i o n will a p p e a r to a g r o u n d o b s e r v e r to fall b e h i n d a s p a c e c r a f t t h a t orbits at less distance. Circular e a r t h orbits at a l t i t u d e s b e t w e e n a b o u t 200 k m a n d 1000 k m are d e s i g n a t e d as l o w e a r t h o r b i t s . Below 200 k m , t h e orbits of s p a c e c r a f t witho u t t h r u s t d e c a y r a p i d l y by a t m o s p h e r i c drag, typically in less t h a n 1 week. Above 1000 km, r a d i a t i o n d a m a g e of u n s h i e l d e d , u n h a r d e n e d electronic c o m p o n e n t s c a n b e c o m e significant. T h e r e is t h e c u r i o u s fact t h a t t h e p e r i o d s of s p a c e c r a f t t h a t orbit close to e i t h e r t h e s u r f a c e of Earth, t h e m o o n , Mercury, Venus, or M a r s are all r o u g h l y t h e s a m e , n a m e l y b e t w e e n 1.4 a n d 1.8 h. This is e x p l a i n e d by the fact t h a t t h e a v e r a g e densities, 0av, of t h e s e celestial b o d i e s do n o t differ greatly f r o m o n e a n o t h e r . It follows f r o m (2.4) for a s p h e r i c a l b o d y of r a d i u s
3.7 Special Orbits R that ix= (4/3)
yrRaGOav,hence
77
from (3.16), with a = R, ~3yr
P =
GOav
i n d e p e n d e n t of the radius.
3.7.2
Geostationary Spacecraft A special case of circular orbits occurs for geostationary satellites. In this case the orbit elements are such that the satellite, for an earth-based observer, remains at a fixed point in the sky. To satisfy this requirement, the orbit must be in the equatorial plane, with the satellite moving from west to east with a siderial period equal to one siderial day ( d s i : 23 h, 56 min, 24.09 s = 86,164.09 s). The radius of the orbit, rgs, therefore follows from
~
]~g
r~s
_
27/"
dsi
with the result that rgs = 42,164.2 km, which is about 6.62 times the m e a n radius of the earth. The discussion here has been limited to the case of a spherically symmetric earth. When taking into account the earth's actual figure, the orbit deviates slightly from the ideal circular orbit in the equatorial plane. The most important effect is an apparent slow change of the geographic latitude of the satellite. This effect, together with other perturbations from ideal Kepler orbits, will be considered in Sect. 5.9.1. At a geostationary satellite's distance from the earth, a substantial portion of the total earth's surface is visible. Three such satellites, spaced 120 ~ apart, cover the entire earth other than regions at high latitudes. For microwave communications, either uplink or downlink, an important consideration concerns the atmospheric and ionospheric effects on signal propagation if the path is too close to the horizon. Particularly in the high range ofmicrowave frequency, such as at K-band (10.9-36.0 GHz), communications at small link elevation angles can be affected by heavy rain, hail, or snow. The principal ionospheric effect at these angles is Faraday rotation. It can be detrimental to linearly polarized microwave links. In practice, it is found that elevation angles above 10 ~ to 15 ~ ensure trouble-flee operations. Figure 3.6 illustrates the earth coverage obtained by three geostationary spacecraft, located 120 ~ apart, assuming a m i n i m u m elevation angle of 15 ~ Each spacecraft covers a spherical earth surface of 154 10 6 k m 2. The gap at the intersection of adjacent regions is at a latitude of 37.4 ~. Earth coverage by geosynchronous satellites, however, involves additional constraints beyond the requirement of m i n i m u m elevation angles. An important consideration is the traffic density, which may differ greatly for different pairs of countries or continents. Important improvements in the utilization of the available microwave spectrum can then be o b t a i n e d
78
C H A P Y E R 3 Orbits and Trajectories in an Inverse Square Field
Figure 3.6 Earth surface covered by three equally spaced geostationary spacecraft, for a microwave link minimum elevation angle of 15~ by shaping the antenna beams, as illustrated in Fig. 3.7 for the Intelsat VI satellite located over the mid-Atlantic. Shaping the antenna b e a m bythe antenna feed makes it possible to concentrate the beam power on the regions where it is most needed and in such a way that the same frequency band can be used for communicating with nonoverlapping regions. Polarization discrimination provides the possibility of an additional doubling of the use of the same band. The advantage gained by placing communications and other satellites into the geostationary orbit has led to a considerable crowding of this orbit. Over the Atlantic Ocean, where geostationary spacecraft handle much of the heavy communications traffic between North America and Europe, the density is particularly high. At some places on the orbit, the spacecraft are separated by less than 1o, forcing the uplinks to operate in different frequency bands and polarizations. The scarcity of the available bands is a major reason for the development of microwave components at ever higher frequency, limited as this approach is, however, by the problem of possible communications dropouts in severe weather. At this time, geostationary spacecraft provide one ofthe most important modes of space communications. They are used in such applications as broadcasting voice and video, telephone service, computer links, and relay satellites for worldwide signal transmissions. An important advantage is that there is no need for slewing of the ground antennas. The apparent small satellite motions resulting from the lack of exact sphericity of the earth can be corrected by pulsing thrusters on the spacecraft so as to keep it within the main lobe of the transmitting and receiving ground antennas. There is unavoidably some delay in the transmissions. This has some effect particularly on telephone connections. The altitude of the satellites above the earth at sea level is about 35,790 km, w h i c h corresponds to a
3.7 Special Orbits
79
Figure 3.7 Atlantic ocean coverage by Intelsat VI spacecraft, with C-band, sixfold frequency reuse. (Courtesy of INTELSAT.) time delay of about 120 ms. Taking into account the lengths of the up- and downlinks including the slant range, together with some (minor) electronic delays, typical one-way communications delays are of the order of a fourth of a second. Other than microwave communications, there are several other applications of geostationary satellites. They include meteorological satellites, earth and ocean observation, and surveillance satellites. Their payloads may operate in m a n y different spectral bands. Again, the large area that can be viewed is often an advantage. On the other hand, the resolution of surface features is necessarily more limited than is the case for low-altitude spacecraft. Also, the requirement for the propulsion to place the satellite into its orbit is more demanding. An important consideration for geostationary spacecraft designed for long-term missions is that the geostationary orbit is beyond the outer Van Allen belt, hence in a region where the radiation environment is relatively benign.
80
C H A P T E Ft 3 Orbits and Trajectories in an Inverse Square Field
3.7.3
Sun-Synchronous Orbits Spacecraft on sun-synchronous orbits have a ground track that crosses a given meridian on the earth once a day ("24-hour satellites"). Also included a m o n g sun-synchronous orbits are those for which the spacecraft crosses a meridian every rational fraction or multiple (e.g., 12, 8, ... or 48, ... h) of a day. Because of the inclination of the ecliptic, different crossings of the meridian through a fixed point on the ground will, in general, have different declinations of the sun. What makes sun-synchronous orbits interesting is that a given point on the earth can be viewed from the spacecraft repeatedly, although not necessarily on each orbit, at the same solar time. This feature is valuable for many tasks of earth observations for which solar illumination is essential. Because the repeated passes occur at similar conditions of solar illumination, such orbits greatly facilitate the correlations of observations of crop growth, drought conditions, changes in the polar ice caps, and other seasonal changes. Sun-synchronous satellites also have great usefulness for communications receivers or transmitters that are located at large latitudes that cannot be reliably reached by links from geostationary satellites. Peak demands for communications and broadcasting tend to occur at the same hours of the day. The orbits are therefore chosen so that at times of peak d e m a n d the satellite is well above the horizon for ground transmitters and receivers. An example is provided by the Russian domestic Molniya system. It is based on orbits with 12-h periods, large eccentricity, and large inclination (Molniya 1993 079A orbit period, 703.1 min; perigee altitude, 436 km; apogee altitude, 39,188 km; inclination to equatorial plane, 62.7~ The apogee is located far north and occurs alternatively over Russia and over North America. Figure 3.8 shows the ground track. Because of the large eccentricity, the time spent by the spacecraft at its apogee, hence at large northerly latitude, is the largest part of the orbital period. A m i n i m u m of two satellites, with ground tracks 12 h apart, are needed to provide continuous service.
O~t
,'
,, .,'
J,/
U "-.
/ - ,
s/ / ..
12,
_ . . -
Figure 3.8 Molniya satellites (Russia) ground track.
%%
o j - -
3.8 Perturbations by Other Astronomical Bodies
81
In principle, a circular satellite orbit in the ecliptic plane could serve as a sun-synchronous orbit. Its period relative to the earth would be 24 h-86,400 s (or a rational multiple thereof); its sidereal period, as obtained from (3.16), in a nonrotating reference frame would be one sidereal day, about 86,164 s. The elevation angle would change in accord with the annual seasons. Because points on the earth with higher latitudes would be poorly covered, such orbits are of only limited interest in comparison with highinclination orbits, such as the Molniya orbits. The orbital planes of Molniya-type sun-synchronous spacecraft must rotate relative to inertial space by one turn per year. This becomes possible by taking advantage of the lack of exact spherical symmetry of the earth's gravitational field. The large inclination of the orbit plane causes the spacecraft motion to be influenced by the zonal coefficients, most strongly by ]2. This is reinforced by the low altitude of the periapsis, where the effects produced by the higher harmonics, which fall off rapidly with distance, are still strong. Even though the perturbations are only modest for a single satellite orbit, they accumulate over m a n y orbits. Of course, these orbits are no longer exact Kepler orbits, as had been assumed otherwise in this chapter.
3.8
Perturbations by Other Astronomical Bodies In this section, motions of spacecraft are considered in which the inverse square force field of an astronomical body is perturbed by another or several other such bodies. Nevertheless, the spacecraft motions will be described, albeit in a very approximate way, by Kepler trajectories. As indicated in Fig. 3.9, a spacecraft at point Ps is in the vicinity of an astronomical body P0. Relative to P0, the spacecraft motion is approximately a conic section; however, it is not exactly so, because not only is the spacecraft motion perturbed by other astronomical bodies P j, j - 1, 2, ..., but also the motion of P0 itself is perturbed by the bodies P j. In some of the applications, P0 may be a planet, with Pj the sun and possibly planets. In other applications P0 may be the moon of a planet and Pj this planet.
Figure 3.9 The derivation of Eq. (3.45).
82
C H A P T E R 3 Orbits and Trajectories in an Inverse Square Field Following (3.1) and referring to Fig. 3.9, the motion of the spacecraft in an inertial reference frame (X, Y, Z) is described by the equation d2Rs
r0__~s
d t 2 = -/z0 r3s
--
Z_,
j-1,2 ....
/zj
rjs rjs
(3.41)
where/zo and ~j are the gravitational parameters ofthe astronomical bodies. Similarly, the mass center of Po satisfies the equation of motion d2Ro _ dt 2 - -
rjo ~
j= l, 2 ....
(3.42)
/ZJ.3
! jO
where ros = Rs - R0 and rjs -- rj0 + r0s. The motion of the spacecraft in a nonrotating reference flame, with its origin attached to the center of mass of Po, is obtained by subtracting (3.42) from (3.41), resulting in
dr2
olros+ro, L
-
r2s
ros
j= l, 2
....
/zo
3 rjs
r,o)] 3 rio
The second term in the square bracket represents the perturbation caused by the bodies Pj. Even though the mass fraction tzj/lzo may be very large (as is the case if one of the Pj is the sun, P0 a planet), this second term is seen to be small c o m p a r e d with unity in two distinct cases, both being of practical interest in astrodynamics: Case 1: If Po is a planet, the P j stand for the sun (and possibly also planets), and ros is very small compared with the dimensions of the solar system (hence r2s is small compared with rgs and rj2o), then rjs and rio are nearly the same and the two terms in the round bracket nearly cancel each other. The same case also arises when a spacecraft orbits the m o o n of a planet at a distance small compared with the p l a n e t - m o o n distance. Case 2: If Po is the sun and P j, j = 1, 2 . . . . , are planets, then it is the mass ratio lzj/tzo that is small. Again, the second term in the square bracket represents a perturbation small compared with unity. In case 1, denoting the magnitude of the spacecraft acceleration caused by Po by aos, and correspondingly the acceleration caused by Pj by ajs, a0s ~
(3.44a)
/zo
r2 s
ajs - /,j
~/1 1 2 c o s qgj LT 4- -a- 2------g--' rjs rio rjsrjo
j -- 1, 2 . . . .
(3.44b)
where from the triangle Ps-Po-Pj and the angles ~j and l~rj indicated in Fig. 3.9,
rjas - rj o + rL - 2rjo os cos cos~j =
rjo - ros cos !kj rjs
3.8 Perturbations by Other Astronomical Bodies
83
Defining the p a r a m e t e r of smallness s = ros/rjo, one obtains
rJ-A
= 1 - 2s cos Oj + s2
rj0
cosCj = (1 - 2s cos ~] + s2) -1/2 (1 - s cos Oj) Substitution into (3.44b) results in the e q u a t i o n ttj
ajs= -jrrio V
1+
1 (1 - 2s cos ~j + S2) 2
2(1 - e cos (1 - 2s cos ~j + 82) 3/2
Expanding by the binomial t h e o r e m the second and third terms in the square bracket and d r o p p i n g all terms of order s 3 a n d higher, we have
ajs--
lz j ros V/1 + 3 cos 2 grj rjo
Hence the relative disturbance acceleration ajs/aos of the spacecraft is given by / \ -ajs = # j [r~ ~3 X/1 + 3 c o s 2 ~ky,
aos
j = 1, 2, ...
(3.45)
tZo \ rio /
D e p e n d i n g on the position of the spacecraft, the m i n i m u m ratio of the disturbance acceleration to the zero-order acceleration is seen to occur for l~rj --- 4-90 ~ and is tzj ( r o s ~ 3
I~o \ r j---o/ The m a x i m u m value is twice this m i n i m u m value a n d occurs w h e n ~ j -- 0 ~ or 180 ~ The relative i m p o r t a n c e of the effective disturbance force is seen to increase with the third power of the satellite distance. Table 3.4 indicates the effects of disturbing astronomical bodies on a geostationary satellite. The ratios ajs/aos of the p e r t u r b i n g to the zero-order (in this case the earth's) gravitational accelerations are listed. These ratios are for the case in which the distance of the disturbing b o d y from the earth is a m i n i m u m a n d ~j = 0 ~ or 180 ~ One concludes that the largest relative disturbances on a near-earth spacecraft are caused by the m o o n a n d the sun. A different application of the d i s t u r b a n c e calculation occurs w h e n a spacecraft orbits the sun [denoted by the subscript ( )hi which is n o w the p r i m a r y b o d y P0, b u t the spacecraft orbit is disturbed by a planet. Here case 2 applies. A calculation analogous to that for case 1 gives the result for the relative disturbance acceleration in case 2 as apl, s ah, s
=
/zplr~,s
(3.46)
/Zhrpl, s
When, on the scale of the solar system, the spacecraft is very close to the planet, case 1 applies. Farther from the planet, case 2 applies. A useful
84
C H A P T E R 3 Orbits and Trajectories in an Inverse Square Field
Table 3.4 Disturbance acceleration of geosynchronous satellites, caused by astronomical bodies, relative to earth gravity Disturbing body
Mass ratio ~*j/Vo
Relative disturbance acceleration ajs / aos
Moon Sun Mercury Venus Mars Jupiter Saturn Uranus Neptune Pluto CentauriA
0.0123 332,946 0.056 0.815 0.107 317.9 95.2 14.6 17.2 0.11 3.6 105
1.7 10-5 0.8 10-5 0.9 10-11 1.1 10-9 0.5 10-1~ 1.2 10-1~ 4.2 10-12 0.6 10-13 1.7 10-14 1.0 10-16 4.0 10-22
Source: Cornelisse, SchSyer, and Wakker. [11] approximation, often sufficient for the preliminary p l a n n i n g of space missions, is t h e n obtained by a s s u m i n g that the two volumes corresponding to cases 1 a n d 2 are separated by a spherical b o u n d a r y of radius rpl centered on the planet. The resulting sphere is referred to as the s p h e r e of influence of the planet. A n u m e r i c a l estimate for this radius can be obtained by equating at the boundary, where rpl,s = rpl, the relative m a g n i t u d e s of the disturbances to each other, therefore
#pl
rh, pl ]
/~h
rpl
from which follows
#pl ~ 2/5 rpl = rh, pl k, ~hh ]
(3.47)
An a p p r o x i m a t i o n that can often be m a d e is to a s s u m e that w h e n the spacecraft is within the sphere of influence of the planet, the spacecraft is subject only to the attraction by the planet and outside it is subject only to the attraction by the sun. Use of this a p p r o x i m a t i o n is therefore referred to as the patched conics m e t h o d . Although case 2 has b e e n discussed in terms of the c o m b i n e d action of the sun a n d a planet, it is often also applicable to the c o m b i n a t i o n of a planet a n d one of its moons. In the case, for instance, of the earth and the m o o n , it is important, however, to note, as indicated in Table 3.4, that the d i s t u r b a n c e s of an earth-orbiting spacecraft by the sun a n d by the m o o n are of similar magnitude. In this case, it is no longer sufficient to consider only
3.9 Planetary Flyby and Gravity Assist
85
Table 3.5 Radii of the spheres of influence of the planets
Planet
rpl in 10 6 k m
Mercury Venus Earth Mars Jupiter Saturn Uranus Neptune Pluto
0.09-0.14 0.61-0.62 0.91-0.94 0.52-0.63 45.9-50.5 51.6-57.7 49.4-54.1 85.7-87.6 11.4-18.8
l'pl as a multiple of the planet's m e a n radius
37-56 101-103 143-147 154-185 648-713 859-961 1940-2130 3410-3490 3560-5870
the magnitudes of the accelerations separately rather than to add t h e m as vector quantities. Because, with the exceptions of Mercury and Pluto, the orbit eccentricities are small, hence rh, pl approximately constant, the radii of the spheres of influence of most planets are also approximately constant, i n d e p e n d e n t of their true anomaly. Table 3.5 lists these radii, corresponding to the minim u m and the m a x i m u m heliocentric distance of the planet. It also lists t h e m in terms of the planet's m e a n radius. The concept of a sphere of influence goes back to Laplace (Laplace 1845), who used it in his studies of comets on trajectories passing close to Jupiter.
3.9
Planetary Flyby and Gravity Assist One speaks of flyby or swing-by w h e n a spacecraft on a locally hyperbolic trajectory passes close to an astronomical body. Among the purposes are scientific observations at close distance and the possibility of obtaining increased spacecraft speed at no expenditure of propellant. Still another purpose may be a desired change in the orientation of the flight path following the flyby. The expression gravity a s s i s t is used to characterize flybys in which the spacecraft energy (in an inertial reference flame) is increased. Flybys passing one or several different planets, or even the same planet more than once, are often scheduled as a part of deep-space missions. In this way, energy increases can be achieved that would be prohibitive if the acceleration had to be obtained exclusively from rocket thrust. Timing becomes critical: the planet, or in other cases several planets, must be in suitable positions in their orbits at the instant of the flyby. To be effective, the spacecraft m u s t pass close to the planet. High accuracy in tracking the spacecraft's position on its course toward the planet is therefore
86
C H A P T E R 3 Orbits and Trajectories in an Inverse Square Field 40~-35 t 30
p~ u
8
I arth
I
I
I
Saturn
E
I
I
I
I
Voyager 2
iter
Neptune
25 20 15
O
10
o
~
I
5
Solar System EscapeVelocity -~
\
I
10
I
I
I
15 20 25 Distance from Sun (AU)
I
30
I
35
I
40
Figure 3.10 Planetary flybys by the Voyager 2 (United States) spacecraft. (Courtesy by NASA.) essential. Typically, one or several midcourse corrections by short periods of thrust are m a d e long before the vicinity of the planet is reached. An example was provided by the U.S. NASA Galileo probe to Jupiter. Two earth gravity assist flybys were used. The second flyby passed the earth at 303.1 km, within I km of the intended altitude. It resulted in a saving of 5 kg of propellant for use at Jupiter. As a further example, Fig. 3.10 illustrates the planetary flybys that were executed on the U.S. Voyager 2 mission. The graph shows the velocity in the heliocentric reference frame as it changed with each flyby. In each case, except in the encounter with Neptune, a velocity increase occurred. Also shown in the figure is the solar system escape velocity as a function of the distance from the sun. As is seen by comparing the two curves, after the flyby of Jupiter the spacecraft already had enough speed to escape the solar system. (The other flybys shown were conducted for the exploration of these planets.) For the purpose of preliminary mission planning, it is sufficient to consider the motion of planet and spacecraft within the planet's sphere of influence. The planet's local orbit can then be approximated by a rectilinear path. (This can be verified by a comparison ofthe planet's acceleration on its orbit with the gravitational acceleration near its surface. Thus, for the earth, the centripetal acceleration that results from the near-circular annual orbit about the sun is only 3.44 10 - 3 times normal gravity.) The path of the spacecraft in a planetocentric reference flame is a hyperbola (Fig. 3.1 la). The semimajor axis is designated by a, the eccentricity by e, and the half-angle between the asymptotes by y. tZpl is the gravitational parameter of the planet and Vl the velocity with which the spacecraft approaches the planet along one of the asymptotes before the path becomes appreciably deflected. Similarly, v2 is the velocity at large distance after leaving the planet. The two velocities have equal magnitudes and we write V = V l =V2
3.9 Planetary Flyby and Gravity Assist
87
+
\\\\
(a)
y
(b)
pl,t=o
Pf~
LPl
/
/ "
s , t = O ~ x
(s)
Vpl
L1 ~,,,, V2~-
-
-
Vpl
(c) Lp 1
PI'
t-0"--..
F i g u r e 3.11 Planetary flyby: (a) spacecraft (S) path relative to planet; (b) in heliocentric reference frame; (c) in planetary reference frame. I_v~,planet's path; L~, spacecraft asymptotic approach path; b, impact parameter. (Dashed lines are projections into the plane of the figure.)
The distance, b, from the approach asymptote to the planet's center is called the i m p a c t p a r a m e t e r , a term adopted from atomic and molecular physics. From (3.14) and (3.10), letting the radius r -+ oo (using the lower sign for hyperbolic trajectories), it follows that a-/~pl/V 2 and 1 + e cos(rr - V) = O, so that e = 1/cos y From the geometric relation shown in the figure, b = ae sin y, hence tan y -
bv2/lZpl
(3.48)
The path of the spacecraft w h e n viewed in the h e l i o c e n t r i c reference frame is more complicated. The velocity in this frame is the vector s u m of
88
C H A P T E R 3 Orbits and Trajectories in an Inverse Square Field
Figure 3.12 Ulysses (European Space Agency) solar mission. After a Jupiter gravity assist, spacecraft flies over the sun's south and north poles. The path is shown (schematically) as it would appear in the heliocentric reference frame.
the planet's velocity (in the general case out of the plane of the planetocentric spacecraft path) and the relative velocity of the spacecraft. In general, the path therefore cannot be embedded in a plane. To determine the path, in practice a numerical integration is required. Nevertheless, simple algebraic relations, to be derived in the following, suffice to calculate from the initial, heliocentric conditions the impact parameter and the change in energy. As an example, the path taken by the Ulysses mission of the European Space Agency is shown in Fig. 3.12. After Earth escape, the path led to a flyby of Jupiter. The result was a large plane change and energy increment that allowed the spacecraft to leave the vicinity of the ecliptic plane and to cross the sun's north and south poles. Figure 3.1 lb shows the paths of the planet and of the spacecraft's asymptotic approach in the heliocentric reference frame. The plane of the figure is a plane parallel to these paths. The separation distance of the paths at the crossover point in this projection is designated by s and the angle between the paths by a. The origin of time is taken as the instant when the spacecraft is at the crossover point. At that time, the planet is assumed to lag behind the spacecraft by a distance l. The initial and final velocities are V~ - Vpl +v~ and I/2 = Vpl + v2. The same projection is also shown in Fig. 3.11c, which represents the planetocentric flame of reference where the planet appears at rest. The vector b, the magnitude ofwhich is the impact parameter, and the final relative velocity vz are shown in projection.
89
3.9 Planetary Flyby and Gravity Assist
In the Cartesian coordinate system (x, y, z) indicated in the figure, the c o m p o n e n t s of the relative velocities are
Vlx -- V1 cos o" - Vpl,
Vlz --
Vly = 111sin a,
- Vp + Vx - 2 Vp, Vl c o s
(3.49)
0! /
The c o m p o n e n t s of the i m p a c t p a r a m e t e r vector can be found from bx = 1 + ~.vlx,
by
=
ZVly,
bz = s
where ~ is a scalar multiple of Vl. Since the scalar p r o d u c t b. Vl = O, it follows that ~. = - l V l x / v 2, h e n c e bx
-- Iv2y/V 2,
b y - -lVlxVly/v 2,
bz- s
In terms of the initial conditions, formulated in the heliocentric reference system, the impact p a r a m e t e r is therefore
l b-
3.9.1
/2V12 sin 2 Vp21q- Vl2 - 2VplV1 c o s a
(3.50)
+ s2
Capture by the Planet The impact p a r a m e t e r is the principal quantity that d e t e r m i n e s w h e t h e r or not the spacecraft will collide with the planet. It follows from (3.10) that the distance b e t w e e n spacecraft a n d planet has a m i n i m u m , r m i n , a t the periapsis and that/'min = h2/(/Apl (1 + e)), where h = by1 = by is the angular m o m e n t u m per unit mass. From (3.48) then follows
b2v2 Fmin --
/-/~pl(1 + e)
,
e--
?
l+tan2g=
~
b2v4
1-t-----g-
~pl
!
Designating the effective planetary radius that m u s t be avoided by Rpl, the condition for collision avoidance then b e c o m e s
Fmin
-
b2v 2 /Zpl [ 1 + v/l-'b(b/ll,pl)2v 4]
> R~I
(3.51)
where v2 and b 2 in terms of the heliocentric initial conditions are obtained from (3.49) a n d (3.50). W h e n the inequality is opposite, the spacecraft will collide or, with application of rocket m o t o r thrust a n d / o r aerobraking, be c a p t u r e d in an orbit about the planet.
3.9.2
Change in Heliocentric Energy The velocity of the spacecraft after the flyby is also easily calculated: v2 has a c o m p o n e n t in the direction of v~ o f - v cos(2y) a n d a c o m p o n e n t in the direction o f b o f - v sin(2F). Therefore v2 - - v cos(2F) Vl/V - v sin(2y)b/b
90
C H A P T E R 3 Orbits a n d Trajectories in an Inverse Square Field or
vbx
Vlx cos(2y)
U2x --"
sin(2y)
vby
(3.52)
V2y = --Vly cos(2y) - --~ sin(2y) v2z = -
vS
b
sin(2y)
The potential energy of the spacecraft before and after the flyby is the same. Therefore the change in energy per unit mass in the heliocentric frame ofreference is (1/2)(V22 - V12). From Vlx -- %1-[- ulx,
Vl y = Vl y,
V2x -- %1-[- l)2x,
W2y -- l)2y,
Vlz = 0
W2z = l)2z
follows 1 (V22 - V12) ~- %1(1)2x - l)lx)
(3.53)
which, after substitution from (3.52) and (3.48), results in
1(V22 2
E
V12) = 2 Vp, COS 2 ~' Vpl
V1 c o s o"
(3.54)
11712 sin2 a~Vp2 l + q2 __ 2VplV1 c o s a ] JZpl
A positive energy gain is therefore obtained if Vpl > V1 c o s (7 --[-
3.9.3
~pl
sin 2 a
(3.55)
1 + V12 - 2 Vpl V1 c o s (9"
Parallel Paths of Spacecraft and Planet Particularly simple solutions are obtained when in the heliocentric reference frame the planet path and approach path of the spacecraft are parallel. In this case, v = Vpl - V~. It then follows from (3.48) that, in terms of heliocentric initial conditions for the flyby, the m i n i m u m distance from the planet's center is b2(Vpl-
V1) 2
rmin~pl-~
~pl2 + b 2 (UP l -- V1)4
> Rpl
(3.56)
The change in energy per unit mass, W, becomes W2-
1 Wl - ~(V2 2 - V12) = 2 V p l ( V p l - V1)[1 + ( b / ~ p l ) 2 ( V p l -
V1)4] -1
corresponding to a relative change
w2Wl
= 4(Vpl/V1)C(Vpl/V1)
1)[1 -[- (b/~pl)2(Vpl - V1)4] -1
(3.57)
3.10 Relativistic Effects
91
Energy is therefore gained by the spacecraft if and only if Vpl > V1. The result also shows that for a given velocity ratio and impact parameter, the gain (or loss) is larger for massive planets. Comparing, for instance, Earth (mean equatorial radius = 6378 km, m e a n orbital velocity = 29.78 km/ s, mass = 5.976 1024 kg) and Mars (3402 km, 24.13 km/s, 6.418 1023 kg) and assuming the spacecraft's heliocentric velocity to be one-half the planet's velocity and the impact parameter to be 1.5 times the planet's radius, one obtains For the Earth flyby:
rmin =
7936 km
(Wz - W~)/W1 = 0.273
For the Mars flyby:
rmin=
4817 km
(We - W1)/W1 = 0.026
For flights to Mercury, Venus flybys are attractive. For flights to the outer planets, Jupiter flybys can provide a major boost to the spacecraft. Earth flybys can also be very useful. Thus, it is possible to design a spacecraft orbit that at its perigee is close to the Earth orbit and at its apogee close to the Mars orbit. Repeated Earth flybys can then be used to rotate the spacecraft's perigee-apogee axis to m a t c h partially the rotation of the Earth-Mars line and thereby provide for the spacecraft to pass repeatedly close to Earth and Mars (although, to provide for reasonably frequent such passes, some a m o u n t of thrust will still be n e e d e d to a u g m e n t the effect of the flybys).
3.10
Relativistic Effects In the applications of astrodynamics to spacecraft, relativistic effects can almost always be neglected. To obtain a measure of the typical magnitudes of these effects, we consider here the m i n i m u m velocity Vest at perihelion that is needed for the escape ot a spacecraft flora the solar system, assuming that the path starts at a distance of 1 AU from the sun. It then follows from (3.18), w h e n applied to the sun, that Vesc = 42.1 km/s. From the relativistic equation for the change in mass m--
m0 x / / 1 - t~2/ c 2
(m = mass, m0 = restmass, v -- velocity in an inertial frame, c = velocity of light in vacuum) follows
m/mo = 1 + 9.87 10 -9 Effects of this magnitude are detectable by high-precision Doppler navigational systems. Although ofgreat interest for testing the theory ofrelativity by space experiments, relativistic effects are of only limited practical interest in today's space technology engineering. In deep-space missions, for instance, relativistic effects that may have accumulated over long periods of time are in practice eliminated by midcourse corrections, which can be based on direct m e a s u r e m e n t s rather than on theory. Because spacecraft are accelerated by gravitational forces, relativity experiments conducted with spacecraft must take into account the effects
92
C H A P T E R 3 Orbits a n d Trajectories in an Inverse Square Field associated not only with special relativity (Einstein 1905) but also with general relativity, which is the generalization of the former and includes gravity (Einstein 1916). As an example, it is of interest to consider a spacecraft on a circular orbit about the earth. By (3.14), the speed of the spacecraft is lower at higher altitudes. By special relativity, a clock on the spacecraft would then run faster than a clock on a lower orbit (but, depending on the orbit radius, generally slower than a comparable clock on the ground). This effect, however, is opposed by the effect of the gravity diminishing with altitude. The result is that at about 3200 km altitude the spacecraft clock runs at the same rate as a clock on the ground. Below this altitude, the spacecraft clock runs slow; above it the clock runs fast in comparison with the ground clock. To improve earth coverage, most navigational satellite systems are deployed at relatively high altitudes where the satellite clocks run at a slightly higher rate than the ground clocks that are used to calibrate the system. In practice, this difference in rate is eliminated by the calibration procedures that are applied to the system. In what follows, we will consider the relativistic correction to Keplerian elliptic orbits. In the general theory of relativity, the dynamic equations are modified so that in place of (3.6) the equation for the orbit becomes [12] d2 u
lZ
dO 2 + u -
- ~ + ot
u2
(3.58)
where, as before, h is a constant of motion, u the reciprocal of the radius, and the constant a is defined by a
=
(3.59)
3/z/c 2
For the sun, with/x = 1.32712 102~ m 3/s z, a = 4430 m. As a is very small compared with astronomical distances, the term a u 2 is m u c h smaller than u and therefore represents a correction term that is added to the classical equation of motion. Equation (3.58) is nonlinear. A standard way to solve it proceeds as follows. Write U--" UO'-I-" Ul
where u 0 - ~ # [1 + e cos(0 - 00)] is the corresponding expression (3.10) for the classical Kepler orbit and Ul the relativistic correction. A close approximation to the solution of (3.58) is obtained by substituting u0 for u on the right-hand side, resulting in the linear equation de Ul
dO----T +
ot lz 2 Ul
--"
=
--h-T-[1 + ecos(0 6t/Z 2
h4 [ 1 + ~
1e 2
-
00)] 2
+2ecos(O-Oo)+~
1e 2
cos2(0-02)]
(3.60)
3.10 RelativisticEffects
93
The trial solution Ul --" Co + ClO
sin(0 - 00) + C2 cos 2(0 - 00)
is seen to satisfy (3.60) with Co=-~--
l+~e 2 ,
0r 2 C~ = --h-7-e,
1 0r 2 e2 C 2 - - - g h--T
so that Ul=-~
1+~
1e2+ e O s i n ( O - O o ) - l e 2 c o s 2 ( O - O o ) ]
(3.61)
As a consequence of the factor 0 in the third term on the right, the amplitude represented by this term increases indefinitely with time. No such increase occurs in the other terms, and their effect on the orbit by comparison remains negligibly small. Dropping these terms and again adding the zero-order term, one obtains U
- - U0 -Jr- U l - - ~
1 + e c o s ( 0 - 00) + ~-a-e0 s i n ( 0 - 00)
or therefore u-~-~
l+ecos
O-Oo-~-TO
(3.62)
which can be verified by expressing the cosine ofthe difference bythe cosine and sine of its parts, expanding these by their power series, and neglecting terms of order (a tz / h2) 2. This final result for the reciprocal of the instantaneous radius of the orbit indicates that the orbit is very nearly a classical ellipse. What is new is that the angular coordinate of the periapsis, hence the orientation in space of the ellipse, slowly rotates with time. In describing the motion of planets (particularly of Mercury, where the effect is most pronounced and more easily detectable), the effect is referred to as the relativistic advance of the
perihelion. For one full rotation ofthe orbit, (a/z / h2)Owould have to increase by 27r. This would require h 2/a/z revolutions of the planet or spacecraft along its orbit, with an elapsed time of (h 2/a/z) P, where P is the Keplerian orbital period as given by (3.I6). The angular rate Aw ofthe advance ofthe perihelion therefore becomes Aw-- 2~ c~/z h2P
(3.63)
As found by Einstein, Aw for the planet Mercury is 43" per century, in agreement with observational data. The relativistic corrections to spacecraft orbits are at present too small to matter for engineering purposes. However, spacecraft could be used for deciding between Einstein's formulation of general relativity and competing theories. Spacecraft have the advantage over astronomical measurements
94
C HA P T E R 3 Orbits and Trajectories in an Inverse Square Field that Aco could be m u c h larger than is the case for Mercury. This is seen from the formula Aa~ =
t~/A1/2
(1 - e 2) a5/2
(3.64)
which follows easily from (3.63), combined with (3.16). The relativistic effect is therefore seen to be more pronounced if the semimajor axis a is small and the eccentricity e close to one. Thus, for an earth satellite with a perigee altitude of 1000 km and apogee altitude above the earth's surface of 5000 km, AoJ = 1.03 10 -12 s -1. Over 10 years, the apogee would be displaced by about 3700 m, a distance that could easily be measured. The practical difficulties, however, are great, because other, far larger effects, such as those caused by the irregular shape and mass distribution of the earth, would have to be taken into account with extreme accuracy.
Nomenclature a
b C e
E
f,g h i M n
P r,R f U
v,V w,W Ol
6 0 lz (.o
S2
( )crc ( )gs
()G ( )h ( )H
semimajor axis (Fig. 3.2); also acceleration [Eq. (3.44)] semiminor axis (Fig. 3.2); also impact parameter (Fig. 3.11) velocity of light in vacuum eccentricity eccentric anomaly (Fig. 3.2) coefficients [Eq. (3.36)] angular m o m e n t u m per unit mass inclination (Fig. 3.4) m e a n anomaly m e a n angular velocity orbital period [Eq. (3.16)] radius radius of sphere of influence reciprocal of radius velocity energy per unit mass right ascension (Fig. 1.3); also constant [Eq. (3.59)] declination (Fig. 1.3) true anomaly (Fig. 3.2) gravitational parameter angles referring to perturbation by third body (Fig. 3.9) angle from ascending node to periapsis (Fig. 3.4); Ao9 angular rate of advance of periapsis right ascension of the ascending node circular geostationary Greenwich sun horizon
Problems
( )p ()pl ()r ()s ( )si ()0
95
periapsis planet radial spacecraft siderial tangential
Problems (1) A geostationary satellite is located above the meridian 230 ~ east of Greenwich. For a ground observer at longitude 202~ ' east of Greenwich and latitude 21 ~19' north (Honolulu), compute the azimuth and elevation angles of the boresight of an antenna directed toward this satellite. (Note that the azimuth angle is conventionally counted starting from north in the easterly direction. The elevation angle is counted starting from the local horizontal.) (2) Consider a Kepler elliptic orbit and on it the two points having true anomalies of 90 ~ and 270 ~ respectively. Show that at these points the (acute) angle formed by the radius vector from the center of attraction and the tangent to the orbit is given simply by cotan-~ e, where e is the eccentricity of the orbit. (3) The ground trace (i.e., the projection downward on the earth's surface, assumed to be spherical) of an earth satellite is assumed to pass through a point with latitude 28 ~ (the U.S. Cape Canaveral launch site). The satellite's projection is assumed to move at an angle of 15~ ' counted from the east and increasing toward the north. Find the m a x i m u m and the m i n i m u m latitudes reached by the ground trace. (4) An earth-orbiting satellite is launched with a perigee altitude (the altitude above the earth's m e a n equatorial radius of 6378.1 km) of 1000 km and apogee altitude of 5000 km. (a) Compute the semimajor axis, the semiminor axis, the eccentricity, and the orbital p e r i o d . Also compute the velocities at perigee and apogee relative to the geocentric, nonrotating reference frame. (b) Set up a computer program to print out the distance from the center of attraction and the true anomaly as functions of the eccentric anomaly at increments of 10 ~ (5)* An earth satellite is on an orbit having a semimajor axis of 40,000 km and eccentricity of 0.150. The time after perigee passage of the satellite is 15,000 s. (a) Compute the eccentric anomaly of the satellite by D'Alembert's method, using the first four terms of the series. Also compute the radial distance from the center of attraction and the true anomaly. (b) Compute the same data from the Fourier-Bessel series, again using the first four terms. Compare the results. (6) Jupiter has an equatorial radius of 71,490 km and a gravitational parameter of 126.71 10 6 k m 3 / s 2. A spacecraft executes a flyby with an impact parameter of 230,000 km. The asymptotic speed of approach of the spacecraft in the planetocentric, nonrotating reference frame is 20.0 km/s.
96
C H A P T E R 3 Orbitsand Trajectories in an Inverse Square Field Utilizing the approximation discussed in Sect. 3.9, c o m p u t e the distance of closest approach from Jupiter's surface. Also compute the turning angle of the spacecraft trajectory. (7)* The following six orbital elements (see Fig. 3.4) of a spacecraft are given: a = 30,000 km, e = 0.500, i = 30 ~ w = 70 ~ ~2 = 150 ~ time of perigee passage tp = 0. An earth ground observer is located at geographic latitude 45 ~ north and at the time of perigee passage has a right ascension of 240 ~ Compute the time relative to perigee passage when the spacecraft first appears above the observer's (theoretical) horizon. Also compute the azimuth of the spacecraft's first appearance as it appears to the observer.
References 1. 2.
3.
4. 5. 6. 7. 8. 9. 10. 11. 12.
Stumpff, K., "Himmelsmechanik," Vol. 1, VEB Deutscher Verlag der Wissenschaften, Berlin, 1959. Poincar6, J.-H., "New Methods of Celestial Mechanics," History of Modern Physics and Astronomy, Vol. 13, translated from the French, American Institute of Physics, New York, 1993. Battin, R. H., 'Tkn Introduction to the Mathematics and Methods of Astrodynamics," American Institute of Aeronautics and Astronautics, Washington, DC, 1987. Baker, R. M. L. and Makemson, M. W., 'Tkstrodynamics," Academic Press, New York, 1960. Danby, J. M. A., "Fundamentals of Celestial Mechanics," 3rd printing, Willmann-Bell, Richmond, VA, 1992. Prussing, J. E. and Conway, B. A., "Orbital Mechanics," Oxford University Press, NewYork, 1993. Wiesel, W. E., "Spaceflight Dynamics," McGraw-Hill, New York, 1989. Gurzadyan, G. A., "Theory of Interplanetary Flights," translated from the Russian, Gordon & Breach Publishers, New York, NJ, 1996. Plummer, H. C., '~_n Introductory Treatise on Dynamical Astronomy", Dover Publications, New York, 1990. Herrick, S., %strodynamics", Vol. 1, Van Nostrand Reinhold, London, 1971. Cornelisse, J.W., Sch6yer, H.F.R., Wakker, K.F., "Rocket Propulsion and Spaceflight Dynamics," Pitman Publishing, London, 1979. Eddington, A., "The Mathematical Theory of Relativity," Cambridge University Press, London, 1923.
4 Chemical Rocket Propulsion Rocket propulsion is based on the principle that a vehicle can gain m o m e n tum by expelling mass stored in the vehicle itself. Rocket propulsion therefore differs, for instance, from the air-breathing propulsion of aircraft jet engines, where most of the mass involved in the propulsion is atmospheric air. In the vacuum of space, a space vehicle can also gain m o m e n t u m by a planetary flyby with gravity assist (Sect. 3.9) or by solar radiation pressure ("solar sailing"). For flight within a planetary atmosphere, such as occurs during the early phase of a launch from the earth's surface, important advantages can also be gained by combining rocket and air-breathing propulsion, possibly in a single engine. Nevertheless, classical rocket propulsion is by far the most important means for propelling space vehicles of all
types. The source of energy in rocket propulsion is almost always the chemical energy released by the combustion of a fuel with an oxidizer. (For a class of small thrusters, chemical energy is released by the decomposition of a single propellant, referred to as a "monopropellant," most often hydrazine). Ion or colloid engines, thermal and magnetoplasmadynamic arc jets, resisto-jets, and related devices use electrical energy for their principal energy source. In space, this energy may be derived from solar cells or conversion from the thermal energy produced by radioisotope sources or by nuclear reactors. Electrical propulsion can have a far larger specific impulse, albeit at the cost of a much smaller thrust, than chemical propulsion. To combine a high specific impulse with a high thrust, rockets with several other modes of propulsion have been proposed. They include nuclear fission thermal rockets in which the propellant, hydrogen, is heated by heat transfer from a reactor. Entirely speculative at this time are rockets that derive their energy from reactions of free radicals or from the combination of para- and ortho-hydrogen, from magnetically or inertially confined fusion, or from antimatter. For comparison, the m a x i m u m theoretical energy release per unit mass for some of these reactions is listed in Table 4.1. None of these reactions, with the possible exception of the fission thermal process, are considered practical today for rocket propulsion. In addition to thrust and specific impulse, there are several other characteristics of rocket motors that are of great practical importance. They include, among others, the weight-to-thrust ratio of the rocket motor, the capability for throttling to adjust the thrust level, restart or multiple pulsing capability, storability of the propellant, nozzle designs that provide an optimal compromise between performance in the atmosphere and in the vacu u m of space, and provision for expanding in space the nozzle exit area ("extendable nozzle"). 97
98
C H A PT E R 4
Chemical RocketPropulsion
Table 4.1 Comparison of the Theoretical Energy
Release of Some Speculative Rocket Energy Sources Source
Free radicals (H + H ~ H2) Nuclear fission Nuclear fusion Matter annihilation
Maximum theoretical energy release (J/kg)
2.20 108 7.12 1013 7.62 10TM 9.0 1016
For comparison, the energy release from the stoichiometric combustion of hydrogen with oxygen, at pressures and temperatures typical of rocket motors, is about 1.50 10 7 J/kg. Approximate values of the specific impulse and thrust of rocket motors and thrusters, when operating in vacuum, are shown in Fig. 4.1. Areas in the diagram that are shaded indicate schematically the current state of the art. Separately shown are (1) chemical propulsion main engines (i.e., the rocket motors of launch vehicles and their boosters, lunar or planetary descent or ascent motors, etc.), (2) chemical thrusters (such as are used for attitude control and station keeping of satellites, some of t h e m making use of monopropeUants, e.g., hydrazine), (3) electric propulsion motors (including ion engines, arc jets, magnetohydrodynamic devices, pulsed Teflon thrusters), and (4) chemical-electric thrusters (e.g., hydrazine thrusters a u g m e n t e d by electric resistance or microwave heating). Points (1) to (12) designate specific motors and thrusters. Significantly different in specific impulse, but not in the achieved thrust, are liquid and solid (including hybrid) chemical propellant motors. As shown in the figure, the regions overlap in terms of the specific impulse. The highest specific impulse is achieved by hydrogen-rich combustion with oxygen. Reactions with beryllium and fluorine can theoretically result in a still higher specific impulse. However, such propellants are too hazardous to be ofpractical value. In principle, high thrust combined with high specific impulse could be obtained with nuclear-thermal rockets. In these motors, hydrogen, heated by heat transfer from a nuclear reactor, serves as the propellant. The predicted performance of the U.S. Nerva project of the 1960s, since abandoned, is indicated by point (8) in the diagram. Electric propulsion comprises ion and colloid propulsion, arc jets, pulsed Teflon thrusters, and various magnetohydrodynamic devices. Ion thrusters have high electric efficiency and represent the most mature technology a m o n g these devices. They are being used on some spacecraft. The area under the dashed curve shows the higher thrusts that could be obtained by clusters of electric thrusters, particularly ion propulsion motors. Still lacking, however, are space-borne nuclear reactors and thermal-electric conversion devices with high enough electric power to drive these thrusters. (The required reactor thermal power in watts is indicated in Fig. 4.1 and is
4.1 Configurations of Liquid-Propellant Chemical Rocket Motors
99
Figure 4.1 Vacuum performance of propulsion systems. Shaded areas are state of the art. (1) Saturn V--Apollo (USA), 5 boosters F-I, RP1/O2 and solid boosters. (2) Ariane V (ESA) at liftoff, Vulcain HM-60 and solid boosters P-230. (3) Proton (Russia), UDMH/N204, 6 boosters. (4) Space Shuttle main engine SSME (USA), H2/O2. (5) DELTA 7925 (USA) at liftoff, RP1/O2 and solid boosters. (6) Ariane V (ESA), Vulcain HM-60, H2[O2. (7) PAM STS STAR 48 (USA), solid motor. (8) Nerva (USA) nuclear-thermal, project abandoned. (9) Resisto-jet, NH3. (10) Pulsed Teflon. (11) MPD arcjet, Ar. (12) Ionbombardment engine, Hg, 30 cm (NASA). b a s e d on an a s s u m e d t h e r m a l - e l e c t r i c c o n v e r s i o n efficiency of 30% a n d an e l e c t r i c - t o - b e a m p o w e r efficiency of 90%.) The t r e a t m e n t in this c h a p t e r has profited f r o m a n d follows to a considerable extent the b o o k by S u t t o n a n d Ross [1] a n d the b o o k by Huzel a n d H u a n g [2].
4 . 1 Configurations of Liquid-Propellant Chemical Rocket Motors A typical d e s i g n of a l i q u i d - p r o p e l l a n t c h e m i c a l m o t o r is illustrated in Fig. 4.2. The fuel, after p a s s i n g a m e t e r i n g valve, enters a c i r c u m f e r e n t i a l m a n i f o l d a n d from there flows t h r o u g h radial p a s s a g e s in the i n j e c t o r plate. The oxidizer, after its valve, e n t e r s a d o m e w h e r e it is d i s t r i b u t e d over the top
100
C H A P $ E R 4 Chemical Rocket Propulsion
Figure 4.2 Liquid-propellant chemical rocket motor. From Ref. 1, Sutton, G. P. and Ross, D. M., "Rocket Propulsion Elements," 5th ed. Copyright 9 1986. Reprinted by permission of John Wiley & Sons, Inc. Courtesy of Rocketdyne Division of Rockwell International, U.S.A. surface of the injector plate. It continues its flow through axial passages in the plate, separate from the radial ones. Fuel and oxidizer are then injected at high pressure through separate, narrow injection holes into the combustion chamber, where they mix and ignite. The resulting combustion gas then accelerates in the nozzle, which consists of a converging (subsonic) and a diverging (supersonic) section. At the t h r o a t where the converging and diverging parts meet, the gas velocity equals its local sonic velocity. The aft end of the nozzle is referred to as the exit plane. The combined structure of combustion c h a m b e r and nozzle is frequently called the t h r u s t chamber. Typical of m a n y large motors, such as the one shown in the figure, is the tubular construction of the thrust chamber. It can be manufactured by brazing together a large n u m b e r of tubes, shaped to result in the configuration required by the nozzle. Before entering the injection plate, all or a portion
4.2 Configurationsof Solid-Propellant Motors
101
of the f u e l ~ or alternatively, but more rarely, the o x i d i z e r ~ flows through the tubes to cool the thrust c h a m b e r walls. Thrust chambers of this type are referred to as being regeneratively cooled. Circumferential stiffening bands around the nozzle are often used, particularly around the throat, where, in the illustrated motor, hydraulic actuators are attached. Their purpose is to provide for steering the vehicle by deflecting the motor structure through small angles in the two orthogonal directions, about 5 ~ to 7 ~ being common. To allow for this motion, the motor is supported on gimbals in mountings that transmit the thrust. The fuel and oxidizer lines must be flexible; stainless steel bellows are often used for this purpose. Although the fuel and oxidizer will usually self-ignite w h e n mixed, a separate pyrotechnic igniter is provided for starting the motor. Its use prevents a dangerous accumulation of unreacted propellant at the start of engine operation. In some cases the igniter is not designed to survive the motor firing and is allowed to be ejected through the nozzle.
4.2
Configurations of Solid-Propellant Motors The principal feature of solid-propellant motors is that fuel and oxidizer, together with a binder and other additives, are alreadypremixed. They are contained in this form in a motor case, rather than in separate tanks as would be the case for liquid propellants. At normal temperature, after curing, the propellant is a solid, rubber-like mass. At a moderately higher temperature it has the consistency of a highly viscous fluid and can be cast into the motor case. Combustion occurs on the inner surface of the propellant and proceeds outward. The geometrical configuration of the propellant is such that the b u r n surface stays roughly constant; this ensures that the gas pressure and thrust stay approximately constant. At the end of the burn, the burn surface reaches the wall of the motor case, which is thermally protected by a noncombustible liner. In casting the propellant into the motor case, a mandrel is used to provide an inner channel, not filled with propellant, to conduct the gas to the nozzle. Often, a second purpose of the mandrel is to form an initial b u r n surface that differs from a strictly circular cylindrical shape. The purpose is to increase the initial b u r n surface, making its area more nearly equal to the area of the final burn surface. Solid-propellant motors are very frequently used as upper stage motors. Often, they also form an integral part of the spacecraft that contains the payload. The motor case, in this case, is usually near-spherical or ellipsoidal in shape, hence structurally nearly optimal to contain the gas pressure. An example of this type of motor is shown in Fig. 4.3. Solid-propellant motors are often the main motors of smaller launch vehicles, such as rocket probes. Very large solid-propellant motors are freq u e n t l y used as boosters, strapped to the sides of major launch vehicles. Boosters operate at least in part in the atmosphere and are therefore given some aerodynamic shaping. This requirement, and the need to conform externally to the main body of the launch vehicle, calls for a solid motor that
102
C H A P T.E R 4
Chemical Rocket Propulsion
m
_
_
v
Figure 4.3 PAM-STS,Star 48 solid-propellant rocket motor. (Courtesy of Thiokol Corporation, U.S.A.) By permission. has the form of a long cylinder. At the end of the burn, the empty motor cases are dropped off. Often, they are recovered to be used again in a later launch. Boosters are often so large that it would no longer be practical to cast the propellant as a single mass. Reasons are the difficulty of casting such masses in the time before they cool and the difficulty of transporting such large objects to the launch site. The propellant and containing motor case are therefore often subdivided into segments that are stacked lengthwise and assembled only at the launch site. Maintaining the integrity of the joints between segments and guarding against gas leaks are critical. O-rings are used in pairs or in triplicates to obtain reliability by redundancy. In turn, they are protected from the heat and chemical attack of the combustion gas by zinc chromate putty or similar materials. Figure 4.4 illustrates this type of booster. Here the major part of the propellant is cast in three cylindrical segments. At the end of the burn, the boosters are severed from the main vehicle by pyrotechnic mechanisms and separated from the vehicle by small rockets or springs. Frequently, large boosters contain a parachute package for recovery of the empty case and nozzle from the ocean. The case, which is either steel or a composite, can then be used after refurbishment for another launch. In comparison with liquid-propellant motors, solid-propellant motors are simpler in concept and construction and lower in cost. Generally, they have a lower specific impulse. In a n u m b e r of applications, a disadvantage of solid-propellant motors is that, once ignited, they produce thrust until all the propellant has been consumed. Because the exact performance of the propellant depends on several factors that are difficult to control to very high precision, the thrust cutoff is less precise than with liquid-propellant motors. To obtain
4.3 Rocket Stages
103
Figure 4.4 Ariane 5, solid-propellant booster. (Courtesy of CNEE France.)
the accuracy required in maneuvers such as orbit insertion, small liquidpropellant thrusters on the spacecraft are needed for correction. (Proposals to terminate the burn at a precise time by explosively removing parts of the nozzle have not found general acceptance.) This disadvantage is overcome in hybrid rocket motors. Here the fuel and oxidizer are stored separately, the former in a solid mass in a configuration similar to that in solid-propellant motors. Cryogenic oxygen, or a gaseous or liquid or gelled oxidizer, is injected into the motor case. Because the oxidizer flow can be precisely controlled, precise thrust termination, thrust throttling and multiple engine starts are possible.
4.3
Rocket Stages The large rockets used for space missions are built and operated in stages, arranged one on top of the others. Strap-on solid-propellant boosters, firing simultaneously with the main liquid-propellant motors, in combination, can also be considered to constitute a single stage, providing thrust from booster ignition to booster termination.
104
C H A P TIE R 4 Chemical Rocket Propulsion By definition, each stage is a complete propulsion system. After completion of the burn, the stages are separated one by one from the remaining stages or payload. The advantage that can be gained by staging follows from the fact that after the separation of a burned-out stage the remaining vehicle is lighter and therefore easier to accelerate to a still higher velocity. As is evident from the rocket equation (2.28), the rocket's initial mass, needed to accelerate a given payload mass to a given velocity, has an exponential dependence on the required velocity increment. The resulting theoretical gain obtained by staging is somewhat reduced, however, by the added mass of the stage separation mechanism and by the unfavorable mass scaling factor of smaller, multiple tanks, feed systems, and motors as contrasted with a larger, single system. For launch vehicles, another advantage of staging is the possibility of designing the motors to perform optimally for the particular atmospheric pressure in which they are principally operating. Stages are usually connected to each other by bolted flanges. Separation is then initiated by the use of pyrotechnic bolt cutters. Another method of connecting stages is by so-called Marman clamps, in which a circumferential steel band presses inverse wedges against pairs of wedges that are parts of the respective two stages. Separation is initiated by cutting the steel band pyrotechnically. The principle of the functioning of these clamps is illustrated in Fig. 4.5. Springs provide a simple means of pushing off the separated stage. To prevent a difficult-to-predict interaction of the separated stage with the rocket motor plume of the next stage, some distance is needed between the separated stage and the remaining vehicle. A second reason for waiting
Figure 4.5 Operating principle of a Marman clamp.
4.3 Rocket Stages
105
for a m i n i m u m distance before firing the next stage is that the separated stage will usually experience a slight residual acceleration caused by gas that evolves from the still hot motor. This acceleration could possibly result in a collision of the separated stage with the remaining vehicle. In place of springs, small rocket motors are also being used. In what follows, an optimal distribution of the masses of the stages of an n-stage rocket will be computed. The calculation, which follows in the main Ref. 3, provides insight into such questions as the optimal sequence of stages that may have different mass, specific impulse, and thrust. It should be noted, however, that most often the designer of a space mission has available only a limited choice of rocket motors, most of t h e m having already been developed for earlier space missions. The calculated, theoretical optimal stage mass distribution therefore can provide only a general and qualitative guide. Let too, i, i = 1, 2, . . . , n, designate the masses, before ignition, of the stages of an n-stage vehicle and ml, i their masses at the end of the burn. Thus m0,1 and m1,1 refer to the first (lowest) and mo, n a n d ml, n to the last (uppermost) stage. The mass of the payload is designated by mpl. (The expression "payload" is frequently used ambiguously; sometimes it refers to the scientific instruments and c o m m u n i c a t i o n systems carried by a spacecraft; in other instances it m a y refer to the total spacecraft. In the present context, "payload" will be understood to be a complete spacecraft, exclusive, however, of main propulsion motors that are sometimes integrated with it; the latter are counted as an additional "stage.") Each stage mass before the b u r n can be regarded as the s u m of its initial, useful propellant mass mpr, i and a residual mass mrs, i. The latter consists of the sum of the masses of the e m p t y propellant tanks or motor case, excess propellant mass carried as a margin of safety, the propellant feed system, thrust chamber, nozzle, actuators, avionics, and stage structural elements. The stage residual mass fraction, ei, is then defined by 6i --
mrs,i / mpr,i
so that mo, i = mpr,i(1 -k-ei) I ml, i ~
I
mrs, i
i = 1, 2 . . . . , n
(4.1)
A convenient concept is that of subvehicles, defined by the s u m of all stages not yet fired plus the payload. Therefore, including the complete launch vehicle and the payload, there are n + 1 such subvehicles. The mass ofthe ith subvehicle before the b u r n ofits lowest stage is designated by Mo, i, and after the b u r n ~ but before separation of the residual m a s s ~ by M l , i. With these definitions n
Mo, i : mpl -k- ~
mpr, j(1 -k- ej)
j=i n
i = 1, 2, ..., n
Ml, i : mpl - mpr, i -+- y ~ mpr, j(1 -+- 8j) j=i
In particular, Mo, 1 is the l a u n c h m a s s of the complete vehicle.
(4.2)
106
C H A PT E R 4 C h e m i c a l R o c k e t P r o p u l s i o n
Useful parameters for characterizing the subvehicle masses are the subv e h i c l e m a s s fraction Pi-- M],i/Mo, i
and the s e q u e n t i a l subvehicle m a s s fraction ~i--
Mo, i+l/Mo, i
In particular, let qJ = mpl/Mo,1, qJ, which is called the payload m a s s fraction, is the most important characteristic for judging the benefit that can be obtained from staging. It follows directly from these definitions that qJ is the product of the sequential mass fractions, that is, n
kI/
: H ~i i=1
(4.3)
The stage masses before the burn can be expressed by the same parameters. It is easily shown that too,1 = Mo,1 (1 - ~1) i-1 mo, i
i -- 2, 3, ..., n
= Mo, l(1 - ~i) H ~J'
(4.4)
j=l
The three parameters because from Vi =
and Ci are not independent of each other,
v~, r
Mo, i -- mpr, i
=1
Mo, i
mpr, i Mo, i
and ~i =
Mo, i - mpr, i - mrs, i
= 1-
Mo, i
mpr, i(1 + Ei) Mo, i
follows ~ i --
vi(1 -I- 8i)
-- 8i,
i -
1, 2 . . . .
, n
(4.5)
Next, we consider the theoretically optimal distribution of stages in two special cases. These suffice to answer, at least qualitatively, most questions related to the general case.
4.3.1
Optimal Distribution; Negligible Gravitational Acceleration It will be assumed that during the period of thrust the gravitational acceleration is negligible compared with the acceleration caused by the thrust. This assumption is approximately valid in m a n y space missions, exceptions being the ascent or descent from a planetary surface and low-thrust, electric propulsion. The thrust is assumed to be parallel to the trajectory, which therefore is rectilinear. The different stages may have different, but constant, specific impulse.
4.3 Rocket Stages
107
Let Ft, i be the magnitude of the thrust of the ith stage. For brevity, we define the characteristic velocity ci - goisp, i. In accordance with the definition (2.24) of the specific impulse and (2.26), (2.27), Ft, i ~ r n i g o i s p , i -~ - c i d M i / d t
- Mi(t) d V i / d t
where M) (t) and 14(t) are the mass and velocity, respectively, of the ith subvehicle. Integrating from the beginning to the end of the burn results in the velocity increment (A V ) i = Vx, i - Vo, i = - c i In vi for the ith subvehicle. The velocity at the start of the burn of subvehicle i + 1 is the same as the velocity at the end of the burn of subvehicle i. Therefore the velocity gain A V of the payload is obtained by s u m m i n g all the separate increments, with the result that n
A V = Vpl-
E c/(ln(1 + 8i)
Vo, 1 =
--
ln(gz/+ ei))
(4.6)
i=1
If the specific impulse, hence the characteristic velocities q, and the stage residual mass fractions are kept constant, the v e l o c i t y g a i n o f t h e p a y l o a d w i l l be m a x i m a l w h e n the (positive) term n
ci In(0/+ ei)
- ~ i=1
has an absolute maximum, subject to the side condition (4.3). Following Vertregt [31, the desired result is obtained by the m e t h o d of Lagrange multipliers. Under mild conditions (not all Jacobians of the side conditions vanish, and certain inequalities among the second derivatives must be satisfied to outrule saddle points) the m e t h o d will calculate the relative extrema. We define the functions n
f(g'1, ~2, . . ., ~/n)
- ~
ci ln(0i + el)
i=1
and n
h(~l, llr2,...,
llrn) --
In qJ - y ~ In l~ri i-1
where h vanishes w h e n the side condition (4.3) is satisfied. There is a constant multiplier, say )~, such that
of
Oh
(4.7)
o~i Evaluating the partial derivatives
of
ci
Ollri
~ i Jr- 8i
Oh
1
C H A P T E R 4 Chemical Rocket Propulsion
108
and substituting t h e m into (4.7), together with the side condition, results in the n + 1 algebraic equations for ~1, ~P2, ..., ~n and )~
i = 1, 2 , . . . , n |
Ci!lri -- ~,(~i + gi),
n ]--[ ~pi = ~
/
(4.8)
i=1
Solving the first of these equations for 1/r i in terms of the Lagrange multiplier gives
~ki =
~.ei ci -
(4.9)
Finally, substitution of 1/ri from (4.9) into the side condition (4.3) results in /1 ~,n ]-'[ ~.
8i = ~ ci - ) ~
(4.10)
which is a polynomial equation of degree n for )v. The m e t h o d of Lagrange multipliers, as discussed, gives the extremal values. To find the absolute m a x i m u m of the function (4.6) it is best, in applications, to examine numerically and separately the real solutions of (4.10). To summarize the entire calculation: After solving (4.10) for the Lagrange multiplier, the values of the sequential subvehicle mass fractions are found from (4.9). The masses of the rocket stages follow from (4.4). Finally, the velocity increment of the payload is found from (4.6). The stage mass distribution that results in the m a x i m u m velocity gain is most directly obtained by comparison of the values of the Lagrange multipliers. A qualitative understanding of the optimal distribution of the stages in a multistage rocket is best obtained by numerical examples. Table 4.2 lists the pertinent data for a two-stage rocket with a payload mass fraction mpl/Mo, 1 of 0.100. One of the stages is assumed to use a solid-propellant motor with a specific impulse of 280 s. The other stage uses a hydrogen-oxygen motor with a specific impulse in vacuum of 455 s. The stage residual mass fraction in both cases is a s s u m e d to be 15%. In case l a, the high specific impulse stage is the second stage to fire; in case ib it is thefirst stage. As noted from the table, the sequential subvehicle mass fractions r are reversed in the two cases and the payload velocity gains A V are the same. The stage mass fractions mo, i/Mo,1, however, differ. Case la is seen to be more advantageous because the hydrogen-oxygen engine, which is more costly, is smaller than in case lb, and this for the same final velocity gain of the payload. This conclusion is not limited to two-stage rockets but is also valid in the more general case of n stages.
4.3.2
Optimal Distribution; Vertical Ascent In the ascent or descent from and to a planetary or lunar surface, the thrust and gravity forces are comparable in magnitude. In the case of large, liquidpropellant launch vehicles, launched vertically from the earth's surface,
4.3 R o c k e t Stages Table 4.2
109
Optimal Stage Distributions
Case 1 Two-stage rocket. Gravitational force neglected compared with thrust. Payload mass fraction ~ = 0.100. Specific impulse Isp,~ = 280 (solid-propellant motor) and 455 s (hydrogen-oxygen motor, vacuum performance), respectively. Stage residual mass fractions e~ = 0.15.
Propellant ~. (m/s) a ~ ~P2 too, ~/Mo, ~
mo,2/Mo,1
First stage Second stage First stage Second stage First stage Second stage
A V (m/s)
Case l a
Case lb
Solid
H2-O2
H2-O2
Solid 2240 0.151 0.662 0.849 0.051 6393
2240 0.662 0.151 0.338 0.562 6393
Case 2
Two-stage rocket. Vertical ascent in constant gravity field. Payload mass fraction ~ = 0.100. Specific impulse Isp, i = 280 and 455 s, respectively. Stage residual mass fractions ei = 0. Thrust-to-gravity fractions q~i = 1.50. Propellant )~ (m/s) a ~1 ~/t2
too, 1/Mo, 1
mo,2/Mo,1 A V (m/s) a
First stage Second stage First stage Second stage First stage Second stage
Case I a
Case l b
Solid
H2-O2
H2-O2
Solid 2472 0.669 0.150 0.331 0.569 4460
2472 0.150 0.669 0.850 0.050 4460
Corresponding to the absolute maximum of A V.
c o m m o n l y u s e d t h r u s t - t o - g r a v i t y ratios at takeoff are a b o u t 1.2 to 2.0 [Space Shuttle, U.S.A., a b o u t 1.42; Proton, Russia, 1.29; Ariane 5, ESA, 2.27]. The principal r e a s o n for c h o o s i n g these relatively small ratios is t h a t for larger ratios the m a s s a n d p o w e r r e q u i r e m e n t s of the l i q u i d - p r o p e l l a n t feed syst e m s b e c o m e excessive as a c o n s e q u e n c e of the very large p r o p e l l a n t flow rates that h e a v y l a u n c h vehicles m u s t handle. For smaller vehicles, s u c h as rocket p r o b e s with s o l i d - p r o p e l l a n t m o tors, this restriction does n o t apply. Relatively high t h r u s t - t o - g r a v i t y ratios at takeoff are n o t only possible b u t also a d v a n t a g e o u s b e c a u s e t h e y minimize the g r a v i t y l o s s (Sect. 2.2.4). W h e r e a s in the p r o b l e m t r e a t e d in the p r e c e d i n g section, time delays b e t w e e n successive firings of the stages do n o t affect the o p t i m i z a t i o n of the stage m a s s distribution, this is n o longer the case here. The effect of delays is m o s t directly seen in the e x t r e m e case of h o v e r i n g as it o c c u r s in the d e s c e n t to a p l a n e t a r y or l u n a r surface w h e n t h r u s t a n d gravitational force
110
C H A PT E R 4
Chemical Rocket Propulsion
just cancel each other. In this case there is no change in either the kinetic or potential energy of the payload, and the entire energy expended by the motor is invested in the gas of the rocket plume. It is evident, therefore, that in the case of motion in a gravitational field it is advantageous to fire the stages without delay. Some exceptions will occur, however, w h e n dictated by atmospheric effects on the performance of different motors. In what follows, we consider the vertical ascent of a multistage rocket in a constant gravitational field. For simplicity, the stage residual masses are neglected and the rocket motor mass flow rates and also the specific impulse of each stage are assumed to be independent of time. Time delays between the firing of successive stages are assumed to be zero. Analogous to the equation of motion in the preceding section, but now adding the gravitational term, ]~(t) dVi/dt = -ci d]V~/dt-
i-
gMi(t),
1,2 . . . . . n
By separation of variables and integration, f
Vl'i
d~ - -ci
fi~ l'i dMi
,I Vo,i
,i
-
g
1~
f
tl, i
dt
dto, i
Since the mass flow rate/~i of each stage is assumed constant, l~ti(tl, i -
tO, i) = M o , i -
M l , i --
Mo,i(1 -
vi)
Also, since the stage residual masses are neglected, M o , i+l -- M l , i,
vi -~ ~ri
Defining the thrust-to-gravity fractions fli :
i~i > 1
Ft, i / ( g M o , i),
and expressing the magnitude of the thrust of the ith stage by Ft, i -- c i l ~ i , where ci is the stage characteristic velocity, to, i = (1
tl,i-
-
(4.11)
~i)r
The velocity increment (A V ) i produced by each firing is therefore given
by ( A V ) i = Vl, i -
Vo, i --
-ci[ln
~r i -[-
(1 -
~ri)/i~i]
Because V0,i+l ~- Vl, i, the velocity increment A V of the payload is obtained by s u m m i n g all stage velocity increments. Hence A V-
Vpl-
Vo,1 -- E ci i=1
In 1 ~i
1 - Oi
(4.12)
/~i
where V0,1 is the initial velocity of the rocket. In a planetocentric reference frame, V0,1 is usually zero, but positive if the rocket is launched from an aircraft. In what follows, we wish to maximize A V by an optimal distribution of the stages, assuming that the characteristic velocities ci and thrust-to-gravity
4.4 I d e a l i z e d M o d e l o f C h e m i c a l R o c k e t Motors fractions
j~i remain
111
fixed. We define the functions
f(grl, ~r2, ..., lPn)- ~
ci(ln 1
i=1
1 - grg)
l/ri
/~i
n
h(~/1, gr2 . . . . .
~ ) - In 9 - y~. In Oi i-1
The theoretically o p t i m u m stage distribution will therefore occur when f has an absolute maximum, subject to the side condition (4.3). Making use of the method of Lagrange multipliers, expressed by (4.7), it follows from the evaluation of the partial derivatives that ~i = fli(1 - )~/ci),
i = 1, 2, . . . , n
(4.13)
Finally, substitution into the side condition results in the nth degree polynomial equation n
I - I fli(1 - ~./ci) i=1
=
*
(4.14)
for the Lagrange multiplier. In turn, the stage masses are obtained from (4.13) and (4.4). The velocity gain of the payload will result from selecting among the real roots of (4.14) the one resulting in the absolute m a x i m u m of AV. Table 4.2, case 2, illustrates the results of such a calculation. As in case 1, a two-stage rocket with a payload ratio of 0.100 is assumed. The advantage of the hydrogen-oxygen motor being the last rather than the first stage (stage mass to initial vehicle mass = 0.050 rather than 0.:331) is even more pronounced than in the zero-gravity case. An alternative way to optimize the stage distribution is to maximize the payload's total energy gain, that is, the sum ofits kinetic and potential energy. Still another example of stage optimization is formulated in the Problems section.
4.4
Idealized Model of Chemical Rocket Motors In this section, certain simplifying assumptions will be made to describe solid- and liquid-propellant rocket motors. With these assumptions, a general understanding of the main features of such motors can be obtained. The theory based on these idealizing assumptions is not sufficient, however, to calculate the performance of such motors to the accuracy required for space mission planning. To obtain this accuracy, recourse must be made to additional theoretical developments the principal ones will be considered in subsequent sections and to testing. The assumptions that will be made presently, roughly in order of their importance, are that (1) the combustion gas satisfies the ideal gas equations and has constant specific heats; (2) the flowthrough the motor is steady state; (3) the velocity immediately following the combustion zone is low compared with the local sonic velocity, so that the thermodynamic conditions there are
112
C H A P T E R 4 Chemical Rocket Propulsion the ones of a stagnation point; (4) the flow properties d e p e n d only on a single spatial variable Cone-dimensional flow"); (5) heat losses to the boundaries and viscous drag are neglected, hence the flowis isentropic, with the possible exception of gas dynamic shocks that may occur in the supersonic region.
4.4.1
Flow Variables Downstream of the combustion zone, conservation of the energy of the gas requires that
i u 2(x) = const. = ho h(x) + -~
(4.15)
where x designates the spatial variable in the direction of the mean flow, h(x) t h e enthalpy, and u(x) t h e velocity of the gas relative to the motor. The subscript ( )0 indicates the stagnation condition. Using conventional t h e r m o d y n a m i c notation, it then follows with h = cpT,
Cp =
Y g-1
R
where ), -- Cp/Cv is the ratio of the specific heats and R the gas constant for the particular ideal gas, that u=
y-1
RTo 1 -
~
(4.16a)
A more commonly used form of this equation is obtained by introducing the pressure in place of the temperature. For the assumed isentropic change of state, = const.
T p -(•215
so that u--
),-1
i
RT0 1 -
P0
(416b)
It is often convenient to use the Mach number, M = u/a, as the indep e n d e n t variable, where a = a(x) = v/? , R T is the local speed of sound. Solving (4.16a) for the temperature then results in 1 T/To - [1 + ~(y - 1)M2] -1
(4.17a)
Making use of the isentropic relation a m o n g the t h e r m o d y n a m i c variables, one obtains similarly
P/Po - [ 1 + ~1 (F - 1) M2 ] - y / ( F - 1 )
(4.17b)
M2] - 1 / ( y - 1 )
(4.17c)
1
0/00 = [1 + ~ ( y - 1) where 0 is the density. Also, from (4.16b)
U=
2F RT0[X - (1 + 2(y 1 - 1) M2 ) -1 ] y-1
(4.17d)
4.4
Idealized Model of Chemical Rocket Motors
113
Finally, the velocity and the t h e r m o d y n a m i c variables can be related to the cross-sectional area, A(x), of the flow by m e a n s of the conservation of mass equation
O(x)u(x)A(x) =
const' =
(4.18)
Taking the logarithm, and then differentiating, results in
do/o + du/u + dA/A
= 0
(4.19)
In turn, from the isentropic relationship between density and temperature, after taking the logarithmic derivative,
do/o
( y - 1) d(T/T~
whereas taking similarly the logarithmic derivative of (4.16a),
du/u =
1 d(T/To) 2 1 - T/To
Solving for d(1"/To), substituting the result into the preceding equation, and using (4.17a) gives for the first term in (4.19)
do/o = -M2 du/u Hence the conservation of mass equation can also be written in the form (1 -
M2)du/u + dA/A
= 0
(4.20)
From this it is seen that to increase the velocity of the flow in the subsonic region (M < 1), the cross-sectional area must decrease, and it m u s t increase in the supersonic region (M > 1). The cross section where M = 1, hence dA/dx = 0, is referred to as the throat. Although it is not directly related to rocket motors, it m a y be observed here that the condition dA/dx = 0 at the throat is merely a necessary but not a sufficient condition for M -- 1, because, as indicated by (4.20), it m a y h a p p e n instead that du/dx = O. Flows that lead from a supersonic to a subsonic velocity, hence the reverse of normal nozzle flows, are also possible. However, they are far less stable and typically include gas dynamic shocks. This type of flow is related to the intake flow of jet engines in supersonic flight. The conditions at the throat are c o m m o n l y designated by an asterisk. Therefore, with this notation, u* - a*. The pressure p* is obtained from
P*/Po =
1
(1 + ~(F --
1))-Y/Y-1
(4.21a)
as follows immediately from (4.17b). In applications to chemical rockets, the effective value of y usually ranges from about 1.20 to 1.40, c o r r e s p o n d i n g to a value of p*/Po of about 0.56 to 0.53. This is the basis for the convenient m n e m o n i c that the pressure at the throat is roughly one-half of the stagnation pressure, i n d e p e n d e n t of all other variables.
114
C H A PT E R 4
Chemical RocketPropulsion
Other useful expressions in terms of the conditions at the throat are, from (4.17c), (4.17d), and (4.18)
1 + ~(---y - I-)M2
O*
)
1/(V-1)
(4.21b)
u ( ~(y+x)M, )1" .-~ = 1 + ~(7 -777,s, A A*
(4.2~c) (v+i)/(•
=
~
(4.21d)
~-(7 + i,i
Figure 4.6 shows the pertinent flow variables, calculated from the preceding equations, for the nozzle contour illustrated in the figure.
m
y = 1.25 Aex/A*
4.0
= 50
_
,
I
0.8
_
3.0 M
u/u*
0.6
_
-
2.0
~
0
r/r 0
-
P/Po
0.4
_
-
1.0
U/U*
TO
j
P/Po
d~o
-
_
0.2
Figure 4.6 Flow variables in a typical rocket nozzle, assuming an ideal gas with ratio of specific heats y = 1.25 and a nozzle area ratio Aex/A* = 50.
Idealized Model of Chemical Rocket Motors
4.4
115
I
i 'k ,
/
I
/"
i
/
i
/
i
11 \ I
I
\ /'\.
!
i
(a)
I
'~
J
,
~
(d)
i
i
(b)
I
\
/
\,
/
|
i I
,
'
,
'\
'
i
/
,
i
\
|
.~
(c) ~ -
~"
J
(e)
Figure 4.7 Schematics of thrust chambers: (a) bell nozzle; (b) extendable nozzle; (c) expansion-deflection nozzle; (d) axially symmetric aerospike nozzle; (e) planar aerospike nozzle. Dashed lines indicate jet boundaries at sea level, dashed-dotted lines those in vacuum. 4.4.2
Nozzle Contours
Configurations in which the cross-sectional area following the combustion zone first decreases, then increases, are typical of all chemical rocket motors. This concept for obtaining a supersonic velocity was first realized by De Laval (1845-1913), who applied it to the nozzles of steam turbines. In rocket motors, a c o m m o n configuration is the bell nozzle, illustrated in Figs. 4.2 through 4.4 and 4.6. The simple, one-dimensional analysis discussed in the preceding section applies equally to other configurations, including extendable nozzles, expansion-deflection nozzles, and aerospike nozzles (when without gas injection, also referred to as ,'plug nozzles"). They are shown schematically in Fig. 4.7. In place of the axially symmetric configurations, planar, twodimensional geometries are also of interest. The aerospike and expansiondeflection nozzles are designed to obtain a more favorable compromise in nozzle efficiency w h e n the motors must operate at changing atmospheric pressures as the launch vehicle rises through the atmosphere. The extendable nozzles are used on medium-sized motors that operate in the vacuum of space. They make possible large ratios of exit plane to throat areas, ratios that are higher than what is feasible for the large motors of launch vehicles that must operate in the atmosphere. The combustion gas jet boundary (shown in the figure schematically by dashed lines for operation at sea level and by dash-dotted lines for vacuu m condition) can vary greatly during ascent. Section 4.6 contains a fuller discussion of the change in nozzle performance as a function of altitude. The aerospike and expansion-deflection nozzles, compared with the classical bell nozzle, tend to be somewhat heavier and present greater complexity in cooling. For upper stage motors operating in a near or complete
116
C H A P T E F1 4
ChemicalRocketPropulsion
vacuum, bell nozzles are therefore preferred. Especially in the case of small motors, the bell shape is often replaced by a straight cone, with little loss of efficiency. As is evident from Fig. 4.7, quite abrupt changes in the nozzle contour near the throat are often employed. The purpose here is to keep the overall length of the nozzle short and to reduce the requirement for cooling. Such contour changes are possible without flow separation because the flow in a normally operating nozzle is strongly accelerating, resulting in a pressure drop that energizes the boundary layer.
4.5
Ideal Thrust A generally applicable expression (2.23) for the thrust of a rocket was obtained in Chap. 2. For the idealized model discussed here, in place of integrating over the exit plane, it suffices to use the m e a n value, Uex, of the gas velocity. Therefore the magnitude of the thrust of the ideal rocket motor becomes Ft : if/Uex -+- ( Pex -- Pa) Aex
(4.22)
The same relation can also be stated in terms of the specific impulse,
1[
Isp - g00 Uex +
(Pex-m-Pa)Aex]
(4.23)
It is noted here that rh, the mass flow rate through the rocket nozzles, in a n u m b e r of cases can be slightly less than the time derivative of the vehicle mass. This occurs with engines that use gas generators to drive the propellant feed turbo pumps, after which the gas is d u m p e d overboard. The result is a reduction of specific impulse, typically by I or 2%. In the normal operation of a rocket motor, the second term on the right of (4.22) or (4.23) is considerably smaller than the first term. This second term is often referred to as the pressure thrust, the first (and larger) one as the velocity thrust. For a given mass flow rate, it follows from (4.20) with M > 1, that the exit plane velocity U~x and hence the thrust become larger as the exit plane area Aex is increased. There is an incentive, therefore, to make this area as large as possible, compatible with weight and size. For rocket motors that operate in the atmosphere, the ultimate size of the exit plane area is also limited by aerodynamic drag. Already m e n t i o n e d have been the extendable exit cones. As indicated schematically in Fig. 4.7b, they consist of one or several nested cones that can be mechanically extended in the axial direction and locked to each other. They have been used successfully in space to increase the exit plane area beyond the limitations imposed b y t h e physical size ofthe rest ofthe vehicle. Theoretically, the m a x i m u m exit plane velocity that could be obtained (with an infinite exit plane area) would be U e x , m a x m_
2F RT0 y-1
4.6 Rocket Motor Operation in the Atmosphere
117
as follows from (4.17d) with M ~ oo. The gain in thrust obtainable from an increased exit plane area can be considerable; for instance, in the case illustrated in Fig. 4.6, the theoretical gain from an infinite exit plane area would be 17%. As seen from this relation, ideally, for a given mass flow rate, the thrust could be increased ifthe combustion temperature (i.e., the propellant chemical reaction energy) could be increased, combined with a large gas constant (i.e., a low molecular weight) and a ratio of the specific heats close to 1. In fact, all these factors are interrelated and cannot be optimized simultaneously.
4.6
Rocket Motor Operation in the Atmosphere When operating in the vacuum of space or in the earth's upper atmosphere, the flow downstream of the nozzle throat is entirely supersonic. Perturbations, as may be induced by the atmospheric back pressure at the exit plane, therefore cannot influence the flow in the nozzle or combustion chamber. In Fig. 4.8, the curve labeled by "a" represents the ratio of the gas pressure to the stagnation pressure under normal conditions (case of "ideal expansion"). The bell nozzle and the flow variables are the same as those in Fig. 4.6. In most applications to launch vehicles, it is sufficient to assume that the back pressure at the exit plane equals the ambient atmospheric pressure. The details of the interaction of the flow of air at the base of the vehicle with the combustion gas at the nozzle exit plane are therefore neglected. If the ambient atmospheric pressure is above a certain limit (which depends on the nozzle geometry and the exit-plane-to-throat area), a gas dynamic shock will form, at first at the exit plane and at still higher ambient pressure in the nozzle interior. Downstream of the shock, the flow will be subsonic. (At still higher ambient pressure, the entire flow would be subsonic, including that at the throat; for rocket motors, this particular condition is of no practical interest, however.) Figure 4.8 illustrates the several cases that can occur, depending on the ambient pressure. Consistent with the previously made assumption of onedimensional flow, the shock, if one occurs, would be normal to the m e a n flow direction. The flow variables, indicated by the subscript ( )1, immediately preceding the shock, and the flow variables, indicated by the subscript ( )2, immediately following the shock are related to each other by the conservation of mass, m o m e n t u m , and total enthalpy (enthalpy plus kinetic energy). For ideal gases, they give rise to a large n u m b e r of alternative, and well-known, equations (e.g., [4]). Some of these are listed in the following for reference. To compute the data shown in Fig. 4.8, the relation P2_( Pl
2y ) M 2 y+l
y-1 y+l
was used. Because for a prescribed nozzle geometry the flow variables ahead of the shock are known from (4.21), and because the ambient atmospheric
118
C H A P T E R 4 Chemical Rocket Propulsion
5-10 -1
2 . 1 0 -1 %
b1
10-1~ 5. I0-2 I
y= 1.25
Aex/A* = 50
o
~,,
%
%
m II
_
2.
%
0 0
Q.
10 - 2 10 -2 _
(-
5.
10. 3 -
r (]J I:L
o (13 .L='
2. 10- 3 -
.o
Z 0_ a,)
b3
x
o
0 Ib41
~
Ideal Expansion
,,
Shock at
o
>o
II tN
(13 r,r r-
~c ["-
0 0
(9
A
b2
n-
I t
k&\\
!
Exit~
10-3 _ _
{::L
"o el/
_
5 . 10 4 -
Vacuum Operation
Figure 4.8 Theoretical pressure ratio P/Po for the nozzle of Fig. 4.6. (a) Ideal expansion; (b~), (bz), (b3), (b4) overexpansion with normal shock; (c) underexpansion; (d) subsonic flow throughout.
pressure is also assumed to be known, the location ofthe shock in the nozzle, as a function of the ambient pressure, can be readily calculated. Thus, the four vertical lines labeled bl to b4 in the figure illustrate the shock pressure rise for four different assumed ambient pressures. The last, b4, represents a shock that is located at the nozzle exit; it connects the pressure of the ideal expansion to the ambient pressure. Other well-known relations that connect the flow variables ahead of and behind the shock are the Hugoniot and Fanno equations. The most generally useful equations in applications to rocket nozzles, however, are those that
4.6
RocketMotorOperationin theAtmosphere
119
use as the i n d e p e n d e n t variable the Mach n u m b e r ahead of the shock. They are =
(
+
2)(2)/ y-1
P2 = 2____~Y^~2 p~ y+l T2 T1
y-1
M1
-1
)-1
;/-1 y+l
- ( 1 + Y-1M12)( 2F 1M12-1) ((Y+1)2M12) -1 2
(4.24)
S2-- Sl _ R
Y ln[ 2 Y-1] -- Y - 1 (y + 1)M12 + y + l +
1
F-1
In
[2y
F+I
M12-
y 1] F+I
where s is the entropy. When the ambient pressure is less than the exit plane pressure corresponding to the ideal expansion, additional expansion takes place downstream of the nozzle. This case is referred to as u n d e r e x p a n s i o n . The reason is that if the exit plane area were larger, additional expansion with additional thrust could be had. In the v a c u u m of space, as a consequence of their finite exit plane area, all rocket motors necessarily operate in the regime of underexpansion. The thrust, as conventionally defined, is given by (4.22). In the second term, the "pressure thrust," the term paAex, which is subtracted from the thrust, simply accounts for the difference of exit plane pressure and ambient pressure at the nose of the vehicle. (Omission of the term would violate the hydrostatic principle that in the absence of thrust the force from a uniform ambient pressure must be zero.) The velocity thrust, r~U~x, for a given exit plane area is unaffected by underexpansion. The pressure thrust (Pen - pa)Aex, in an u n d e r e x p a n d e d flow with a fixed exit plane area is seen to be increased in v a c u u m c o m p a r e d with operation in the atmosphere. If the ambient pressure is larger than the ideal expansion exit plane pressure, there will be a gas dynamic shock, first at the exit plane and at still higher ambient pressures in the nozzle interior (Fig. 4.8). This condition is referred to as overexpansion. Since behind a normal shock the velocity is subsonic, there will be a further reduction of the velocity in the divergent section of the nozzle and at the exit plane, where the gas pressure matches approximately the ambient pressure. The result will be a diminished thrust. In case there is a shock inside the nozzle, U~xin the formula is no longer the value obtained from the ideal expansion. Also, the stagnation pressure downstream of the shock differs from that upstream. The exit plane velocity in this case is obtained by using the shock relations (4.24), followed by the ideal compression of the subsonic flow as discussed in Sect. 4.4.1. In Fig. 4.9, the thrust, Ft, of a motor operating in the atmosphere is compared with the thrust, Ft, vac, of the same rocket motor in vacuum. The ratio Ft/Ft,vacis shown as a function ofthe ratio of the combustion c h a m b e r
120
C H A PT E R 4
Chemical Rocket Propulsion
1.00 0.98 0.94 0.90
Ideal Nozzle, No Losses
_
F~
J _
,, "
0.62
-."7
.. - --..~.". ~ . ~ - ~ ' ~ "
/"
-o/ /
/
Nozzle Operating Range for Aerospike Nozzle %___._~
y= 1.25
A ex/A* = 50
'
~
-m ~
i 200
Aerospike Nozzle
s
/
7~-uF
_ ~ I/o 0.74 - ~
0.66
,-'f
S
Ft, vac 0.78
0.70
~
2 .2 -
086-/I ,' 0.82
~/-
Space Vacuum
Operating Range for Bell Nozzle i
I
I
500
I
I
I
I I
1000 Po / Pa
I
2000
I
I
I
5000
Figure 4.9 Performance comparison of comparable bell, aerospike, and ideal nozzles in the earth's atmosphere (the ideal nozzle as in Figs. 4.6 and 4.8). stagnation pressure P0 to the a m b i e n t pressure Pa. The data are for the same values of the effective ratio of the specific heats, y = 1.25, and of the area ratio Aex/A* = 50 that were used to construct Figs. 4.6 a n d 4.8. The curve labeled "ideal nozzle" represents the case of o n e - d i m e n s i o n a l flow of an ideal gas w i t h o u t losses.
4.6.1
Comparison of Different Types of Nozzles Including all losses, typical thrust ratios of aerospike and of bell nozzles are shown in Fig. 4.9 for c o m p a r i s o n with the ideal nozzle. The improved perform a n c e in the a t m o s p h e r e of the aerospike (and, similarly, of the expansiondeflection nozzle) in c o m p a r i s o n with the bell nozzle is evident. Also indicated in Fig. 4.9 is the p e r f o r m a n c e at sea level. In this case, an u p s t r e a m stagnation pressure of 2200 N / c m 2 (the c o m b u s t i o n c h a m b e r pressure of the U.S. Space Shuttle main engines) is assumed. It m a y be noted for c o m p a r i s o n that if the same rocket m o t o r a n d nozzle were operated in the a t m o s p h e r e of Mars (ambient atmospheric pressure at Mars m e a n radius 4.9 m b a r - 490 N / m 2, hence a b o u t 0.5% of the pressure at earth sea level), the thrust of a bell nozzle would be close to 99% of the v a c u u m thrust. W h e n bell nozzles are used in launch vehicles, the throat-to-exit-plane area ratio needs to be chosen as a c o m p r o m i s e b e t w e e n w h a t would be optimal for p e r f o r m a n c e in the lower a t m o s p h e r e and optimal for the higher a t m o s p h e r e or in vacuum. A t ~ e earth's sea level, for the ideal bell nozzle of Figs. 4.6 a n d 4.8 and a c o m b u s t i o n c h a m b e r stagnation pressure of 2200 N / c m 2, the thrust would
Rocket Motor Operation in the Atmosphere
4.6
Toroidalc h Annularthroat
a
m b ///I~I
e
~
121
Primaryflow actson nozzle, ~\ producingthrust
Nozzle ~ Innerfree-jet boundary Outerfree-jet boundary -~ Trailingshock
1
/ \~-~~/ II
~,~
/<,,~ - - Secondaryflow actson base, \ producingthrust Primaryflow Subsonic recirculatingflow
Figure 4.10 Flow field of an aerospike nozzle in the atmosphere. From Ref. 2, Huzel, D. K. et al., "Modern Engineering for the Design of Liquid Propellant Rocket Engines." CourtesyRocketdyne Division of Rockwell International. Copyright 9 1992, AJAAm reprinted with permission. be 86% of the vacuum thrust. At the density scale height of the atmosphere (i.e., at 9290 m altitude), the thrust would already be 96% of the vacuum thrust. Again taking as an example the Space Shuttle main engines, their nozzle exit pressure is about 0.08 atm. They do not approach the ideal expansion condition until an altitude of 18 km is reached. The loss of thrust at sea level compared with vacuum is about 20%. This loss of thrust of the main engines is tolerable because the solid-propellant boosters (which are designed for low-altitude performance) provide 80% of the liftoff thrust. Aerospike nozzles have a more complex flow pattern, as indicated schematically in Fig. 4.10 for the case of an axially symmetric configuration. The combustion gas exits from the annular combustion chamber and throat to form a jet that is b o u n d e d partially by the surface of the spike and partially by the separation surface formed with the atmospheric air. The deflection of the combustion gas by the spike produces Mach compression cones that then coalesce into an oblique shock cone. The pressure of the gas on the spike contributes substantially to the total thrust. Depending on altitude and vehicle velocity, the flow pattern changes and adapts to the variable air pressure. Because the combustion chamber and the throat have a large diameter, the aerospike nozzle tends to be somewhat heavier than a corresponding bell nozzle with its smaller throat diameter. Nevertheless, the aerospike nozzle is often preferred because of its superior adaptability to different atmospheric pressures. Truncation of the aerospike saves weight but produces a wake region with some attendant losses. With proper design, this loss can be kept small. Expansion-deflection nozzles (Fig. 4.7) have configurations that are intermediate between those of bell and aerospike nozzles. The flow tends to close quickly behind the plug, creating a low-pressure wake. This reduces the thrust significantly, unless either gas from the t u r b o - p u m p system or air is bled into the wake.
122
C H A P T E R 4 Chemical Rocket Propulsion Solid-Propellant Rocket Motor
Nested Cones
Extended Cones \,,
\ ~"
Extension Mechanism Nozzle Locks
Figure 4.11 Solid-propellant rocket motor with extendable exit cone. (Courtesy of United Technologies Chemical Systems Division, U.S.A.) E x t e n d a b l e e x i t c o n e s (Fig. 4.11) have already been mentioned. They
are suitable for ablatively or radiatively cooled nozzles; regenerative cooling as illustrated in Fig. 4.2 is evidently not practical. Engines that are designed exclusively for operating in space in principle would require testing in a vacuum chamber. Because of the practical impossibility of providing ground test facilities with the required pumping rates, large space engines therefore must be tested at atmospheric conditions. If the nozzles designed for space operations were used in such tests, the nozzles might operate in a strongly overexpanded mode with internal shocks and flow separation. It may then be necessary to ground test the engines with nozzles that have an exit-plane-to-throat area ratio smaller than the design value and to rely on analytical methods to predict the thrust that will be obtained in space.
4.7 4.7.1
Two- and Three-Dimensional Effects Nozzle Internal Shocks The occurrence of internal shocks in the nozzle at high overexpansion was already mentioned in Sect. 4.4.4. The discussion there, however, was only qualitative. In fact, rather than a normal shock, there tends to be flow separation, coupled to one or more oblique shocks. Wakes at approximately ambient pressure will be formed downstream of the line of separation so that the nozzle will no longer flow full (Fig. 4.12). The flow will be unsteady, with a separation line that oscillates and with shocks and wakes that often rotate. The engines cannot be operated in this range of overexpansion,
4.7 Two- and Three-Dimensional Effects
FlowSeparation~ . Superson,c~
~
123
~-Subsonic I\~ ~ C \ ~ \
Wake Typical
e
Figure 4.12 Schematic of strongly overexpanded bell nozzle with flow separation and oblique shock. not only because of the drastic loss in thrust but also because of the often destructive engine vibrations that are induced by the nonsteady flow. An empirical relation, applicable to bell nozzles, states that flow separation in the nozzle is likely to occur when
ot(pa/ Po) > (Pex/ Po)
(4.25)
where a varies from about 0.25 to 0.35. This condition is known as the Summerfield criterion. The range for which it applies at typical operating conditions is indicated in Fig. 4.8. Of lesser practical significance are the two- and three-dimensional flow patterns downstream of the nozzle exit plane. If the flow is underexpanded, the plume will contain a series of Mach cones (the characteristics of the hyperbolic equation that governs the flow) that interact with each other and reflect at the gas-air boundary. In the overexpanded case, oblique shocks, combined with expansion and compression Mach cones, occur that interact with each other and reflect at the jet boundary, forming an approximately diamond-shaped pattern. The gas in the plume may have cooled to the point where, because of partial condensation of the combustion gas, the shocks become visible. 4.7.2
Plume Deflection
At or near t h e v a c u u m condition, parts of the plume near the nozzle exit may be so strongly deflected that they impinge on nearby surfaces of the spacecraft. The resulting contamination by the combustion gas can increase the absorption of solar radiation, thereby raising the temperature of the spacecraft. Often more critical is the contamination by the plume of the surfaces of optical scientific instruments and of spacecraft c o m p o n e n t s such as earth sensors. An estimate of the effect can be obtained by using well-known results from the continuum, supersonic flow of the turning of a gas at the edge of a boundary. Figure 4.13 illustrates this case for an assumed nozzle exit plane Mach n u m b e r of 2.00 and a ratio of the specific heats of 1.25. When
124
C H A PT E R
Chemical Rocket Propulsion
4
Stream Lines Mach Lines
.~
/
/" /
I
~
/
I
I
I
I
I
I
Nozzle
\
\
\
M=2.00
3.00
/
4.00
Figure 4 . 1 3 Backflow of rocket plume with potential contamination of spacecraft surfaces. Figure is drawn for Mex = 2.00, y = 1.25. exhausted into vacuum, s o m e small fraction of the nozzle flow can be deflected by more than 90 ~. The fan of lines e m a n a t i n g from the edge of the nozzle are lines of constant Mach n u m b e r (and therefore also of constant pressure, density, etc.). Also indicated are three typical streamlines. Only the streamlines that issue from the nozzle at points very close to the nozzle b o u n d a r y can reach nearby spacecraft surfaces. The angles formed b e t w e e n streamlines and lines of constant Mach n u m b e r are the Mach angles # = sin-l(1/M)
(4.26)
The angle 0 t h r o u g h which the flow turns, starting from 0ex at the nozzle exit plane, is given by 0 - Oex = v ( M ) - v(Mex)
(4.27)
where v(M), for t w o - d i m e n s i o n a l flow, is the Prandtl-Meyer function vCM) = f
=
~ / U 2 1 dM 1 + r-~M 2 M
g-1
tan -1
(M 2 - 1) - t a n -1 v/M 2 - 1
g+l
(4.28)
The constant of integration has been chosen such that at M = 1, v = 0. It follows that the m a x i m u m value of v is Vma~ = g
•
1
1
(4.29)
The c o r r e s p o n d i n g m a x i m u m possible turning angle, given by (4.27), in addition d e p e n d s on the exit plane Mach number.
4.7 Two- and Three-Dimensional Effects
125
Conversely, for a given turning angle, the Mach n u m b e r follows from (4.27) and (4.28). Important for estimates of the level of contamination from the gas flow is the density, which can be calculated from the same relation (4.2 lb) that was derived for the isentropic flow through the nozzle.
4.7.3
Thrust Correction for the Nozzle Divergence Angle In principle, nozzles could be designed that have a parallel, uniform gas flow at the nozzle exit. Such designs are used in supersonic wind tunnels. They apply the method of characteristics to the supersonic flow of the gas, usually assumed to be ideal, to construct the wall contour. For rocket propulsion, the theoretical advantage would be that the lateral velocity components at the nozzle exit would be zero. These components do not contribute to the thrust, yet they subtract from the available energy. In practice, one deviates somewhat from this optimal design. In most bell nozzles, the nozzle contour at the exit plane makes a small angle, called the divergence angle, with respect to the axis of symmetry. Typical divergence angles range from 10 to 18 ~ There are several reasons for this type of design. One is that nozzles with positive divergence will be shorter and less heavy. Without divergence, a short nozzle would require a pronounced concave curvature next to the exit plane, which could lead to nozzle internal shocks and resultant losses. Also, in solid-propellant motors, a strong concave curvature could lead to excessive erosion of the wall from the impingement of liquid droplets or solid particles, such as a l u m i n u m oxide, which are swept downstream by the gas stream. Shorter nozzles with positive divergence also have the advantage of a lower total heat flux into the nozzle structure. This translates into a reduced requirement for cooling. For these reasons, a design compromise is usually made by allowing a small, nonzero divergence angle. The resulting loss of thrust is small. A simple estimate can be obtained by assuming that the flow of the combustion gas can be approximated by a spherically symmetric, inviscid flow that is tangential to the wall at the nozzle exit. The magnitude of the exit velocity and the thermodynamic quantities such as pressure and density are therefore assumed to be constant on the spherical cap that spans the nozzle exit. Let rex the radius of the cap, Sex its area, ~ the polar angle, and flex the divergence angle. Then Sex = 2Zrre 2 f0 ~x sin ~bd4b - 2zr rex 2 (1 - cos flex) The mass flow rate is now given by r h - 2Jr rexOexUex(1 2 - cos flex)
(4.30)
Designating the magnitude of the thrust by F(, it follows from conservation of m o m e n t u m that in place of (4.22)
F; -
f ex
COS, dm +
f ( Pex - Pa) COS* aSex ex
(4.31)
C H A P Y EFt 4 Chemical Rocket Propulsion
126
where
L/
r
Uex c o s , d m = 2Jr reax/~ex0ex
ex
sin
~bcos ~bdt
Jrreax/~ex0exsin2 flex
d 0
r~ Uexsin 2 ~ex 2 (1 - c o s jSex)
r~ Uex(1 + cos/Sex) 2
The second integral in (4.31) represents the pressure thrust, which, because it is already small in comparison with the velocity thrust, does not need correction. Hence the loss in thrust is given by F t -- F t =
l#///ex(1
-- COS ]Sex)
(4.32)
For example, as this equation indicates, the relative loss of thrust is 0.8% for flex = 10 ~ and 2.4% for flex = 18 ~
4.8
Critique of the Ideal Model The most important deficiency of the ideal gas model considered in the preceding sections is the assumption of a thermally and calorically perfect gas. In fact, the composition of the gas, following the combustion, can change greatly as a consequence of the chemical reactions a m o n g the various species present. Partial dissociation in the combustion zone and partial recombination of the dissociated species in the downstream parts of the nozzle also play an important role. As the gas rapidly expands, not only will there be a shift in the chemical composition, but also, because the density in the h i g h - M a c h - n u m b e r section of the nozzle is low, chemical equilibrium itself may be lacking. For these reasons, thermochemical calculations of rocket motors are often performed for two extreme cases that are intended to bracket the true process: one is a calculation based on the assumption of a shifting equilibrium, that is, a local chemical and t h e r m o d y n a m i c equilibrium that changes from station to station in the nozzle; the other is a calculation based on the assumption of frozen species, that is, the assumption that the composition attained after combustion remains the same throughout the expansion. Another important p h e n o m e n o n is the excitation, at high temperature, of the vibrational modes of the molecules. One consequence is that the ratio of the specific heats can be quite different from its value at normal temperature. A result of statistical mechanics shows that
n+2 F=
n
where n is the n u m b e r of degrees of freedom of the molecule. Vibrational excitation provides an additional degree of freedom, hence F decreases. As is evident from (4.16b) when applied to the velocity, U~x, at the nozzle exit, even a small change of Y can result in a substantial change in Uex and hence in the thrust.
4.9 Elements of Chemical Kinetics
127
Even if chemical equilibrium may be lacking, it is still possible to define at each location in the thrust chamber a single temperature for all species present. This is because, as a consequence of the high combustion chamber pressure at which m o d e r n rocket motors operate, the molecular collision rates are sufficiently high for quick thermalization to take place. Less critical than the assumption of an ideal gas is the neglect, in the preceding sections, of the heat lost to the walls of the thrust chamber. An estimate of this quantity is essential for judging the structural integrity of the chamber. It directly influences the required rate of cooling or, in uncooled nozzles, the rate of ablation. However, the energy lost in this way by the gas is typically quite small in comparison with the total enthalpy (enthalpy plus kinetic energy). The assumption of a constant total enthalpy that was m a d e in (4.15), therefore, is usually well satisfied. An exception occurs only in the case of very small motors, particularly those that are pulsed intermittently. Such motors are used for attitude control and station keeping of spacecraft. Their large surface-to-volume ratio and the fact that the walls are cool at the start of each pulse substantially reduce the temperature of the gas. In some cases, the time-averaged specific impulse for a pulse may be lowered from the theoretical value by as m u c h as a factor of 2. All newly developed rocket motors go through a long series of test firings. In addition to ensuring the structural integrity of the motor, the tests serve to check and improve the theoretical predictions of the thrust and related quantities. Measurements taken for this purpose include among others the thrust, mass flow rate, combustion chamber pressure and temperature, nozzle throat temperature, and nozzle exit total and static pressures. Small and medium-size motors can be fired on a test stand in an evacuated or so as to simulate the atmospheric pressure at various altitudes a partially evacuated chamber. But for large motors, the p u m p i n g requirements needed to maintain an approximate vacuum would be so high that testing in a chamber is no longer feasible. These motors therefore m u s t be tested in the open atmosphere. It then becomes necessary to infer the thrust that would occur in vacuum or at altitude by the type of calculations that were discussed in Sect. 4.4.4. Testing, as necessary as it is, of course does not obviate the need for a theoretical approach. Engineering calculations must go hand in h a n d with the design. They can eliminate early, less than optimal designs and can greatly reduce the n u m b e r of tests that are ultimately needed.
4.9
Elements of Chemical Kinetics To describe with some precision the processes of combustion of rocket propellants and of the subsequent expansion of the reaction products in the nozzle, it is necessary to apply the methods of chemical kinetics [5, 6]. It is beyond the scope of this book, however, to develop this topic to the point that would be needed to carry out calculations in detail. Only a general survey is provided, sufficient to give the reader some appreciation of the work done by specialists.
128
C H A P T E R 4 Chemical Rocket Propulsion
4.9.1
Chemical Thermodynamics of Ideal Gases The basic principles that govern the properties of the gas after combustion as it expands in the nozzle are the same for all liquid-propellant and most solid-propellant motors. The gas in the nozzle can be described as a mixture of chemical species that is h o m o g e n e o u s at any particular station in the nozzle. As the chemical species can react with each other, the composition of the mixture varies as it flows through the nozzle. The species referred to can be molecules, atoms, or ions. In some cases, rotationally a n d / o r vibrationally excited states are taken into account as separate species. Sometimes more than a single phase of a species can occur. This happens in solid-propellant motors that contain a l u m i n u m as a part of the fuel. In these motors, liquid droplets and solid particles of a l u m i n u m oxide occur. This and other similar cases will not be covered in the following discussion. Because of the relatively high pressures at which rocket motors operate, it can usually be assumed that at a specified station in the nozzle there exists a single temperature c o m m o n to all species. Also, this temperature is sufficiently high so that the laws governing a thermally perfect (but generally not calorically perfect) gas apply. Therefore, if p is the pressure, 0 the mass density, T the absolute temperature, R the gas "constant," (not really a "constant" here since it depends on the changing composition of the gas) and e the internal energy per unit mass, all referring to the m i x t u r e at some point in the nozzle,
p- oRT}
(4.33)
e= f(T)
As indicated by the second equation, the internal energy of an ideal gas or a mixture of ideal gases is a function of the temperature. As a consequence, the enthalpy, h, and the specific heats co at constant volume and Cp at constant pressure of the mixture are also functions of the temperature. Let ha, n2. . . . be the numbers in mole of the species S1, $2,... present in a mixture. It will generally be more convenient to express the quantity of each species by its m o l e fraction, defined by
Xl
nl /
i= 1
Eli,
X2
rt2]
i=1
Xk--
Fli ,
nk
ni
(4.34)
i=1
Hence k Xi--1 Z i=1
The partial p r e s s u r e is defined as the pressure that each gas in the mixture would have if it occupied the whole volume alone at the same temperature. The partial pressures are
Pl
= X1
p,
P2 = X2 p,
...,
Pt = Xkp
(4.35)
The sum of the partial pressures is therefore equal to the pressure p of the mixture (Dalton's law).
4.9 Elements of Chemical Kinetics
129
Similarly, the entropy of a mixture of ideal gases is the s u m of the partial entropies (Gibbs' theorem). It can be shown that the entropy s of the mixture is given by
s=
Tf
dT cp T
R In p Po
(4.36)
The internal energy e and enthalpy h of the mixture are
e=
cv dT,
h-
cpdT
(4.37)
In most applications, including those in this text, only differences ofs, e, and h matter. Therefore the constants To and P0 are arbitrary, although they are frequently fixed at 25~ and 1 atm. The quantity by mass of species Sj in the mixture is expressed by the mass fraction gj, j = 1, 2, ..., k. In accordance with (4.34), gj = lzjnj
(4.38)
Is i=1
where ~j is the molecular weight of species Sj. It follows that k
~gj-1 j=l
The specific heats and the gas constant of the mixture, w h e n expressed in terms of the corresponding quantities for the separate species, become k
- Egjc ,j, j-1
4.9.2
k
k
Egjc ,j, j=l
n- Egjnj
(4.39)
j-1
Degree of Reaction Up to this point, the species in the mixtures were assumed to be nonreacting. In what follows, we introduce reactions among them, but as before, the discussion is limited to homogeneous mixtures that consist of singlephase species and that satisfy the ideal gas laws. When discussing reactions, the species in the mixture are often referred to as the constituents of the reaction. Reactions are represented by the chemical equation "" 9 u _ 2 S - 2 --[- v - i S - 1
<
> 1)1S1 -'['- v2S2 J r - ' "
(4.40)
The reaction may proceed from left to right, in which case the left-side constituents are referred to as reactants and the right-side ones as products. Or, the reaction may proceed from right to left, with the designations for reactants and products reversed. The same species may appear on both sides (which then makes the distinction between a reactant and a product moot). Whether the reaction proceeds to the right or to the left, and how far, will depend on the temperature and pressure (or any other independent variables of state) of the mixture. In applications to rocket engines, the constituents are primarily molecules in their ground state or vibrationally and rotationally excited. However, in describing the intermediate
130
C H A P T E R 4 C h e m i c a l Rocket Propulsion
steps in such reactions, free radicals and atoms often need to be taken into account. The stoichiometric coefficients, v j - . . . 2, 1, 1, 2, . . . , indicate how the reactants can combine into products, restricted by the requirement that the n u m b e r ofeach kind ofatom needs to be conserved. Without any change in the meaning of (4.40), the stoichiometric coefficients can each be multiplied with the same arbitrary constant. A convenient convention that will also be used here is to choose the multiplicative constant such that the stoichiometric coefficients assume their smallest, yet integer value. The n u m b e r s ofmoles of constituents Sj present inside a closed volume at some specified stage in the reaction will be designated by the mole n u m bers nj, as defined before. In contrast to the stoichiometric coefficients, the mole n u m b e r s will be different at the initial, intermediate, and final states of a reaction. In general, they will d e p e n d on the variables of state, such as temperature and pressure. Additional complications (not considered here) arise w h e n chemical and thermodynamic equilibrium is lacking. Mole n u m b e r s can also apply to species that do not participate in the reaction, yet are present in the mixture. An example occurs in rocket motors that use liquid hydrogen and oxygen for their propellant. These motors are frequently operated hydrogen rich, that is, with a hydrogen/oxygen mixture ratio that exceeds the stoichiometric one. This lowers the molecular weight of the combustion gas, with the result that the specific impulse generally will be higher (Sect. 4.4.3). If N is the initial mole n u m b e r of oxygen and Ne the excess mole n u m b e r of hydrogen, and assuming that the reaction goes to completion, the combustion gas will contain 2 N moles of water vapor, Ne moles of hydrogen, and no oxygen. Knowing the specific heats and the gas constant of the constituents, the corresponding quantities for the combustion gas then follow from (4.39). In place of mole numbers, it is often more convenient to use the mole fractions defined in (4.34). The degree of reaction, E, is a measure of how far a reaction, either from left to right or right to left in (4.40), has gone in producing a specified product species. This particular species will be designated by $1 if the reaction of interest goes to the right and by S_1 if to the left. The degree of reaction is defined such that ~ = 0 before any of the specified product has been m a d e and e = 1 when at least one of the reactants is exhausted (so that no additional a m o u n t of this product can be made). If e is known, the mole fractions of all other constituents in the reaction can be found from their stoichiometric coefficients. Illustrated in Table 4.3 is the calculation for the dissociation of water vapor, at high temperature. Hydrogen is taken as the specified product S~; N~ is the final n u m b e r ofmoles of hydrogen w h e n the dissociation is complete.
4.9.3
Equilibrium and the Law of Mass Action A mixture is in thermal and chemical equilibrium w h e n at a fixed temperature and pressure there is no change in the a m o u n t s of the constituents. This is to be understood at the macroscopic level; on a microscopic scale, there will in general be chemical transformations in both directions, as
131
4.9 Elementsof Chemical Kinetics
Table 4.3 Dissociation of Water Vapor: Mole Fractions as Functions of the Degree of Reaction
Constituent
5-1
--
HaO
Stoichiometric
Numbers of
Mole
coefficients u
moles n
fractions x
P-1
=
2
n_l = N1 (1 - e)
n_l
X-1
~ n --
--
nl $1 = H 2
$2 =O2
Vl = 2
v2 = 1
nl = Nle
1
n2 : ~Nle
Xl =
x2 = ~ n -
+ E/2 E
1 +e/2
2., n n2
l-e
1
_
e/2 1 +e/2
E n = NI (1+ 2)
symbolically indicated in chemical e q u a t i o n s such as (4.40). At equilibrium, at the m a c r o s c o p i c level, these t r a n s f o r m a t i o n s will cancel each other. I t can be s h o w n that there will be equilibrium if the Gibbs function G = H - TS (H = enthalpy, T = t e m p e r a t u r e , S = e n t r o p y of the system) has a m i n i m u m as a function of the degree of reaction. If the rates of a reaction are very high in c o m p a r i s o n with the changes of t e m p e r a t u r e a n d pressure, there will be an equilibrium at each m o m e n t . This equilibrium will generally differ from instant to instant. I m p o r t a n t examples are the gas flows in the nozzles of certain rocket motors. W h e n the a p p r o x i m a t i o n o f i n s t a n t a n e o u s , local equilibrium at each station in the nozzle is valid, the process is referred to as one of shifting
equilibrium. In other cases, a better a p p r o x i m a t i o n to the nozzle flow is o b t a i n e d by a s s u m i n g that d o w n s t r e a m of the c o m b u s t i o n zone r e c o m b i n a t i o n is sufficiently slow that there is no further c h a n g e of the constituents. The process t h e n is referred to as one of frozen equilibrium. The law o f m a s s a c t i o n , s t a t e d next, a s s u m e s that the mixture is one of ideal gases in equilibrium. As can be shown, at a specified t e m p e r a t u r e there exists a so-called e q u i l i b r i u m c o n s t a n t , K (T), which is characteristic ofthe particular reaction being considered. The "constant" in this expression refers to the fact that K (T) is independent of the pressure. It depends, however, on the type of reaction a n d is a strong function of t e m p e r a t u r e . The law reads as follows: .
.
.
.
-
K(T)
(4.41)
a n d is valid for an arbitrary n u m b e r of constituents, with mole fractions ... x_2, x_i, Xx, x 2 . . . a n d stoichiometric coefficients ... v_2, v_i, vx, v2 . . . . The pressure of the gas mixture is designated here by p, whereas the symbol Po refers to a s t a n d a r d pressure, usually t a k e n as i atm. [Clearly, if a different s t a n d a r d pressure, say p~, were adopted, this would merely multiply K by (Po/P'o) to the power v_2, v_i, vx, v2 . . . . ] The m a s s action law applies only w h e n there is equilibrium.
132
C H A PT E R 4 Chemical Rocket Propulsion Interchanging the algebraic signs of the subscripts results in the law of mass action for the reaction in the reverse direction. Therefore the two equilibrium constants are related by K1 ( T ) K _ I ( T )
(4.42)
= 1
If the equilibrium c o n s t a n t is k n o w n for a specified reaction and temperature, the mass action law provides the additional information needed to determine the degree of reaction and from this the mole ratios of the constituents in the mixture. For instance, for the dissociation of water vapor, substituting into the law the mole fractions calculated in Table 4.3, one obtains e
e/2
1 - e
1 + e/2
1 + e/2
1 + e/2
p
- K(T)
hence 83
p
(2 + e)(1 - e) 2 Po
= K (T)
(4.43)
In principle, the equilibrium constants could be d e t e r m i n e d by quant u m mechanics. Except in the simplest cases m which are of little interest for rocket propulsion ~ such calculations are far too difficult and time consuming, even with the m o s t advanced c o m p u t e r s available today. Semiclassical m e t h o d s are being used, but for m o s t rocket propellants even these m e t h o d s are of only limited applicability because of their limited accuracy. Equilibrium constants therefore are d e t e r m i n e d experimentally. The classical experimental m e t h o d makes use of flow tubes filled with a gas mixture of known composition. In this tube the gas is heated and flows sufficiently slowly that equilibrium is established at the higher temperature. The gas then enters a capillary tube, where it is cooled very rapidly, in effect "freezing" the higher t e m p e r a t u r e equilibrium composition. The mole ratios corresponding to the frozen equilibrium are then d e t e r m i n e d by chemical analysis. As an example, values of the equilibrium constants for the dissociation of hydrogen and oxygen are listed in Table 4.4 as functions of the temperature. Mso listed is the degree of reaction at a pressure of the mixture of 1 atm. In Table 4.5 are listed the mass action laws for several reactions a m o n g two and three constituents, designated by the letters A, B, C, D. Here the mass action laws are expressed in terms of the degree of reaction rather t h a n in terms of the mole fractions as in (4.41). M1 of these equations, and similar ones for more complicated reactions, can be o b t a i n e d by the m e t h o d that was used to derive (4.43) for the dissociation of water vapor. Given the equilibrium c o n s t a n t and the pressure, the degree of reaction is the only u n k n o w n a n d can be obtained as a solution of these polynomial equations. Having found the degree of reaction, the mole fractions are found in the same m a n n e r as in Table 4.3 for the dissociation of water vapor. Table 4.5 also d e m o n s t r a t e s an i m p o r t a n t physical principle: If the s u m s of the n u m b e r s of moles on the two sides of the reaction are the same,
4.9
Table 4.4
Elements of Chemical Kinetics
133
E q u i l i b r i u m C o n s t a n t s for the D i s s o c i a t i o n
of Hydrogen and of Oxygen and Degrees of Reaction at 1 atm H2 < T(K)
>2H
log K
>2 0
02 <
e
log K
e
5000
1.655
.958
1.712
.964
4000
0.449
.642
0.373
.610
3500
-0.410
.298
-0.580
.248
3000
-1.547
.839 10 -1
-1.854
.590 10 -1 .286 10 -1
2800
-2.115
.438 10 -1
-2.481
2600
-2.767
.207 10 -1
-3.215
.123 10 -1
2400
-3.528
.861 10 -2
-4.071
.461 10 -2
2200
-4.425
.307 10 -2
-5.081
.144 10 -2
2000
-5.496
.893 10 -3
-6.289
.360 10 -3
the degree of reaction An increased
at a given temperature
is i n d e p e n d e n t
(decreased)
pressure
drives the reaction
the fewer (larger) number
ofmoles,
the more so when
of the pressure.
toward
the side with
this relative difference
is l a r g e .
Table 4.5
The M a s s Action Laws for Various C h e m i a l R e a c t i o n s
Reaction
K ( T) =
Reaction
K ( T) =
~C+2D
; C
e 1- e
2A<
A <
e3 ( p / p ~ (2 + e)(1 - e) 2
A <
.~ 2C
4e2(p/Po)
3A <
.~ C + D
e2(3 - e) 27(1 - e)3(p/Po)
3A <
; C + 2D
3A<
~2C+2D
1 - e2 A < ; 3C
27e3(P/p~ (1 -- e)(1 d- 2e) 2
2A<
3A <
A <
; C
~C
; C-I-D
A < ~C + 2D
e(2-e) 4(1 - e)2(p/Po) e(3 - 2e) 2 27(1 - e)3(p/p0) 2
e 2(P/Po) 1 - e2
4e3(P/p~
.~ 2 C + 2 D
16e4(p/po) 3 (1 - e)(1 + 3e) 3
A+B<
A+B
e(2-
~C
"
~. C + D
4(1 - e 2)
4e2
~ 2C
A + 2B <
~C
A+2B
~ 2C
e)
e)2(p/Po)
(1 -
(1
--
,5') 2
e(3 - 2e) 2 4(1 - e)3(p/po) 2 e2(3 -- e)
<
e2 2A<
16e4(p/P~ 27(3 + e)(1 - e) 3
(1 - e)(1 + 2e) 2 A<
4e3 27(1 -- e) 3
(1 -
e)3(p/Po)
e(42A+2B
<
~C
16(1 -
3e) 3
e)4(p/p0) 3
A b b r e v i a t e d f r o m M. W. Z e m a n s k y , i n A m e r i c a n I n s t i t u t e o f P h y s i c s H a n d b o o k , C h a p . 4, M c - G r a w Hill, N e w York.
134
C H A P T E R 4 Chemical Rocket Propulsion
4.9.4
Reaction Rates In the process of expansion in the nozzles of rocket motors, the chemical composition of the gas will differ depending on the station in the nozzle. The expansion can be so rapid that depending on the type of propellant, pressure, and t e m p e r a t u r e ~ c o m p l e t e chemical equilibrium is not achieved, not even locally. The flow properties and the performance of the motor then depend on the reaction rates that govern the chemical reactions. In these reactions there will almost always be intermediate species, for instance, free radicals, that participate. The speed with which chemical equilibrium is reached will then depend on m a n y separate rates for the interaction of these species. The seemingly simple recombination of hydrogen and oxygen to water in fact is the result of a chain of reactions. Omitting various excited states from the scheme, the most important steps are as follows:
H +02
> OH + O
O 4- H2
> OH + H
OH + H2
> H20 + H
H +02 +X
> HO2 + X
H+H+X H+OH+X
(4.44)
> H2+X > H20+X
The reactions are seen to be initiated in this case by atomic hydrogen and oxygen and by the formation of OH radicals. (X is an arbitrary species that can serve as a collision partner in a reaction.) There can take place simultaneously c o n s e c u t i v e reactions and c o n c u r r e n t reactions. Many reaction rates of the intermediate species in chemical rocket motors are only poorly known. Chemical kinetics, although a very old science with a very large literature, is still underdeveloped in terms of engineering needs. The theory is not yet developed well enough to calculate reaction rates a priori [6]. Laboratory methods are limited to relatively narrow ranges of temperature and pressure. Engineering applications must then rely on theoretical extrapolations. Especially with new propellants, it is not u n c o m m o n that some of the reaction rates of the intermediate species are not known within factors of 3 or more. When all of the more important rates for the formation of the species, including the intermediate ones, are known at least approximately, a system of differential equations can be formulated for the numerical computation of the overall rate of the reaction. Because the individual rates in a chain can differ from each other by many orders of magnitude, there may be reactions that are particularly slow, which then result in "bottlenecks" in one or more paths of the reaction. If this is the case, simplifications in the system of differentialequations are often possible.
4.10 Chemical Kinetics Applications to Rocket Motors
4.10
135
Chemical Kinetics Applications to Rocket Motors In the theoretical treatment of the reactions and expansion of the gas, starting from the mixture of propellants in gaseous form in the combustion zone and ending at the nozzle exit, one considers a (small) volume containing a fixed mass of gas moving downstream. The concentrations of the mixture's constituents as defined by (4.48) change because of the reactions among them. In general, there could still be other causes for the concentrations to change. In the presence of gradients, there can be diffusion of the various constituents. In the very small rocket motors used for spacecraft attitude control or station keeping, diffusion to the walls with subsequent catalytic recombination on the walls can play an important role. In the m u c h larger motors of launch vehicles and upper stage vehicles, diffusion in the core of the flow can be neglected, although it plays an important role in the boundary layer, where it affects the heat transfer to the wall. For calculating performance parameters of large motors, such as the specific impulse, the heat lost to the walls by conduction and radiation and the losses due to shear in the boundary layer are negligible compared with the thermal energy of the gas. For these purposes, isentropic, steady-state flow can be assumed. Most calculations [7, 8] are based on the assumption that the gas properties and velocity depend only on a single spatial variable ("one-dimensional flow"). Improvements obtained by including three-dimensional effects would in most cases be insignificant compared with the uncertainties produced by chemical kinetics effects. If important reaction rates are not known or only poorly known, the performance can at least be bracketed by two cases: one is to assume that the chemical equilibrium among the constituents, once established in the combustion zone, remains unchanged in the subsequent expansion ("frozen equilibrium"). The other is to assume that chemical equilibrium is locally established at each station in the nozzle, although the equilibrium composition is allowed to vary from station to station ("shifting equilibrium").
Frozen Equilibrium Calculations The basic notions are the same as those described in Sect. 4.4. The only
difference is in dropping the assumption that the specific heats can be taken to be constant. In rocket motors, because of their high gas temperatures, various vibrational levels of the molecules can be excited. For this reason, the temperature dependence of the specific heats needs to be taken into account in all but the most coarse calculations. As a consequence of the high temperature, or else low density, of the gas in rocket motor nozzles, the gas mixture can be assumed to a good approximation to be thermally (but not calorically) perfect. It therefore satisfies the ideal gas equation. The specific heats cv(T) and cp(T) of the gas mixture are computed from the mole n u m b e r s and the constituents' specific heats from (4.38) and (4.39). In general, the latter must be found experimentally, with only partial
136
C H A P T E R 4 Chemical Rocket Propulsion support from thermodynamic theory. The internal energy and enthalpy of t h e mixture are calculated as functions of the temperature from (4.37). The gas constant, R, of the mixture is also calculated from (4.39). Because the composition of the mixture, by assumption, remains the same, R = cp(T) - c ~ ( T ) is constant throughout the flow in the nozzle. Based on these approximations, the equilibrium flow in a rocket motor nozzle can be obtained from the set of equations (4.45a) to (4.45c) listed next. It is convenient to choose as dependent variables the temperature T(x), pressure p(x), and flow velocity u(x) as functions of the spatial coordinate x. Because viscous losses can be neglected and the flow is assumed to be one-dimensional and steady state, Euler's equation in fluid mechanics applies in the form u d u + l dp = 0 Q
Expressing the density O(x) by the perfect gas law, this becomes RT udu + ~dp0 (4.45a) P Since the flow is assumed to be adiabatic, conservation of energy requires
cp(T) dT + u d u = 0
(4.45b)
With A(x) designating the cross-sectional area of the nozzle, conservation of mass (the "continuity equation" in fluid mechanics) is expressed by O(x)u(x)A(x) - const. Taking the logarithmic derivative, do 0
du u
-F ~
-F
dA ~4
-- 0
If the density is again expressed through the perfect gas law, this becomes dT T
dp p
du u
=
dA A
(4.45c)
If the geometry of the nozzle, hence A(x) and d A / d x , is prescribed, all thermodynamic variables of state and the flow velocity can be computed from these equations as functions of the spatial variable x. The initial conditions that are usually imposed are those in the combustion chamber. The gas velocity there is quite low, usually no more than 10 to 20 m/s, resulting in a very low Mach number. In effect, the conditions there can be equated approximately with the flow's stagnation condition. Following the combustion, equilibrium can usually be assumed in the combustion chamber. Distinct from the idealized rocket model developed in Sect. 4.4, it is now no longer possible to obtain algebraic relations for the various quantities of interest. Instead, a numerical integration of the set of equations (4.45) is required. The results are often expressed by stating them in terms of the temperature and pressure in the combustion chamber and of the expansion ratio of the nozzle or else the pressure at the nozzle exit plane (which usually differs
137
4.10 Chemical Kinetics Applications to Rocket Motors
from the ambient pressure, as discussed in Sect. 4.4.4). Alternatively, the results m a y be stated in terms of the ambient pressure. If the ratio of throat to exit plane areas is such that the thrust per mass flow rate is a m a x i m u m at the ambient pressure, it is then c u s t o m a r y to refer to the specific impulse as the specific impulse for the optimum expansion at this pressure. If operating in vacuum, or near vacuum, the specific impulse is referred to as the
vacuum specific impulse.
Shifting Equilibrium Calculations Because the gas composition varies as a function of the spatial variable, there is now one additional d e p e n d e n t variable. Chosen for this variable will be the degree of reaction e(x). The mole fractions of the chemical species that are present will vary from station to station in the nozzle. The calculation that was s u m m a r i z e d in Table 4.3 for the dissociation ofwater vapor can easily be expanded to the general case. If, as before, the degree of reaction is defined as referring to the product species $1, with stoichiometric coefficient Vl and mole n u m b e r Na within a closed surface, it follows that the mole n u m b e r s of the various species are given by ni :
n-i : N1 (v_i/Vl)(1 - s),
N1 (1)i/1)l)e,
i : 1, 2, 3 , . . .
and therefore the mole fractions by xi -- ni / ~-~ n,
X-i-"
n_i/~n
where j -- 1, 2, 3, ... I)1
y
y
Hence v_i(1 - e)
Vie 3(4
--
~ j vj + (1 - s) ~ j v_j
,
X- i --
e Z j vj + (z - e) ~ j v_j' i = 1, 2 , 3 , . . .
(4.46)
If the specific heats for the individual constituents are known as functions of the temperature, the specific heats for the gas mixture, co(T, e) a n d cp(T, e) and also the gas constant R(e) can be obtained from (4.38) and (4.39) as functions of the degree of reaction. Since, as before, viscous losses are neglected, and the flow is a s s u m e d to be one-dimensional, Euler's equation in the form RT u d u + --:-dp = 0 I - '
(4.47a)
applies, with the only difference that the variables of state now d e p e n d explicitly on e.
158
C H A PT E R 4
Chemical Rocket Propulsion
C o n s e r v a t i o n of e n e r g y is e x p r e s s e d by
(4.47b)
cp(T,e)dT + udu = 0 a n d c o n s e r v a t i o n of m a s s , as in (4.45c), by
dT
dp
du
T
P
u
=
dA
(4.47r
A
T h e law of m a s s a c t i o n p r o v i d e s t h e final e q u a t i o n t h a t is n e e d e d . As in (4.41), w i t h t h e d e p e n d e n c e of t h e m o l e f r a c t i o n s o n t h e d e g r e e of r e a c t i o n n o w m a d e explicit,
(
Xl(__.~!2 c 2 ( ~ ) v 2
"""
)(~00)
v l + v 2 + . . . . (V_l +!)_2+... )
X_l (S)V-l X_2(S) v-'2 ..
= K (T)
(4.47d)
O p t i m u m Mixing Ratio, Combustion Temperature, and Theoretical Specific Impulse of Liquid Propellants
Table 4.6
Pc = 1000 psi a to v a c u u m , w i t h A ~ , / A * = 40
Pc = 1000 psi a to Pex = I a t m . Oxidizer
rm, opt
Tc (K)
Isp(S)
rm, opt
Tc (K)
Isp(S)
4.13 3.21 2.89 2.38 2.58 0.92
2758 3278 3338 3504 3421 3149
389 310 307 312 300 313
4.83 3.45 3.10 2.59 2.77 0.98
2996 3308 3369 3539 3446 3164
455 369 366 371 358 353
N2H4
7.94 2.32
3707 4479
412 365
9.74 2.37
4003 4486
479 430
N204
MMH c N2H4 BsH9
2.17 1.36 3.18
3140 3010 3696
289 292 299
2.37 1.42 3.26
3143 3011 3724
342 344 359
IRFNAa
MMH c UDMH e
2.43 2.95
2971 3001
280 277
2.58 3.12
2965 2995
331 329
MMH r
3.46 2.05
2738 2669
285 287
3.69 2.12
2725 2663
337 338
1.16 1.27
2249 2459
341 327
1.16 1.27
2249 2459
403 390
02
Fuel
H2 CH4 C2H6 C2H4 RP-1 b N2H4
F2
H202
H2
N2H4 N2H4
B2H6 BsH9
Calculations based on shifting equilibrium of isentropic, one-dimensional expansion. a 1000 psi = 689.5 N / c m 2. b "Rocket Propellant-1." c Monomethylhydrazine. a Inhibited red fuming nitric acid. e Unsymmetric dimethylhydrazine. From Rocketdyne Division of Rockwell International, 1992 [2].
4.11 Liquid Propellants
139
Table 4.7 TheoreticalSpecific Impulse of Solid and Hybrid Propellants. Expansion to Sea Level Pressure a Propellant Solids PBAN, 18% A1, 70% NH4C104 DB, HMX, 20% A1, NH4 CIO4 DB, 16% Be, NH4C104 PBA, 20% A1H3, 65% NO2CIO4 Hybrids Bell2, 81% F2 H2, 02, 15% A1 H2, 02, 29% Li H2, 02, 26% Be
Isp 265 270 280 290 382 390 405 456
Expansion from Pc = 1000 psi (= 689.5 N/cm 2) to Pex = 1 atm. From Sutton and Ross [1].
a
Substituting here for the xi the expressions found in (4.46), the law of mass action provides at each step of the integration a polynomial e q u a t i o n for e as a function of p a n d T. The complete set of e q u a t i o n s (4.47), together with (4.46), provides the basis for a step by step n u m e r i c a l integration of the system of differential equations for finding 7", 17, u, a n d e as functions of the spatial variable x.
4.10.1
Sample Results for Several Propellants Calculations of the theoretical specific impulse of m a n y propellant combinations have b e e n carried out and published. In the United States they are largely based on JANAF t h e r m o p h y s i c a l data. Most calculations are based on the a s s u m p t i o n of either frozen or shifting equilibrium. Tables 4.6 a n d 4.7 contain data, based on shifting equilibrium calculations, for selected liquid, solid, and hybrid propellants. The data in Table 4.6 for liquid propellants apply to mixing ratios (mass of oxidizer to mass of fuel) for which the theoretical specific impulse is a m a x i m u m . Separately listed are the p e r f o r m a n c e at sea level a n d in v a c u u m (the latter for a nozzle exit plane to throat area of 40). The data in Table 4.7 for solid and hybrid propellants apply to sea level conditions.
4.11
Liquid Propellants All rocket propellants need to be h a n d l e d with extreme care. Oxidizers in contact with combustible materials of all types can cause u n c o n t r o l l e d fires. In contact with the skin, they can cause severe burns. Fuels are similarly hazardous. The c o m b u s t i o n p r o d u c t s of m o s t propellants are toxic.
140
( H A P T E R 4 Chemical Rocket Propulsion Other than low-thrust electric thrusters (and, potentially, nuclear rocket motors), liquid-propellant rocket motors generally have the highest specific impulse. They can be designed for very high thrust, as needed for launch vehicles, but can also satisfy the low-thrust requirements of spacecraft attitude control and station keeping. Excluding some propellants such as liquid fluorine (an oxidizer) and beryllium (a fuel, combined with liquid hydrogen), which are extremely hazardous, the highest specific impulse can be obtained by the combination of liquid oxygen and liquid hydrogen. An excess of hydrogen above the stoichiometric ratio lowers the molecular weight of the combustion gas, hence increases the specific impulse. As an example, the main engines of the U.S. Space Shuttle use a mixture ratio by mass of oxygen to hydrogen of 6.0 with a nominal specific impulse of 363 s at sea level. The most c o m m o n liquid propellant systems use a separate oxidizer and fuel. Such propellants are referred to as bipropellants. Distinct from them are monopropellants, such as hydrazine (N2 H4), which release their chemical energy by decomposition, usually by means of a catalyzer. Monopropellants have the advantage of requiring only a single tank and feed but generally have a lower specific impulse. They are used principally on spacecraft. Propellant combinations that ignite at normal temperatures without the need for an igniter are called hypergolic. A useful distinction is also made between cryogenic propellants, such as liquid oxygen and hydrogen, and noncryogenic ones. Propellants may be earth storable, that is, may be stored for prolonged periods at normal temperature, or space storable. The latter are usually defined as having a boiling point at 1 atm higher than 123 K. In selecting suitable propellants, a n u m b e r of other considerations are important. They include their fire and explosion hazards, toxicity of the boiloff, and corrosion of materials that are in contact with them. Several of the most c o m m o n oxidizers, fuels, and monopropellants are described in the following. A more comprehensive list of their physical properties is found in Appendix D. 4.11.1
Oxidizers
Liquid oxygen, which boils at 90.0 I( at 1 atm, is one of the most frequently used oxidizers. It is noncorrosive and, except for evaporation losses, can be stored indefinitely. When in contact with organic materials at normal pressure, it usually will not cause spontaneous combustion. However, at higher pressures, in the presence of lubricants and other inflammable materials, it can cause a violent explosion and fire. In contact with the skin, it can produce severe burns. As with all cryogens, it is necessary to insulate thermally tanks, fluid lines, and valves to reduce the loss from evaporation. When used in large quantities, such as for launch vehicles, liquid oxygen is often economically produced directly at the launch site. Nitrogen tetroxide (N204) at normal temperature and pressure is a yellow-brown liquid that produces reddish-brown, highly toxic fumes. In containers made from compatible materials, it is storable at normal
4.11 Liquid Propellants
141
temperature for prolonged periods. However, w h e n not protected from contact with air, it will readily absorb moisture and form a strong, corrosive acid. It is hypergolic with m a n y fuels. When in contact with organic materials, for instance wood, it will cause spontaneous ignition. Nitrogen tetroxide is frequently used as the oxidizer on launch vehicles but also on spacecraft as one of the propellants in bipropellant systems. Its relatively high freezing point o f - 12~ necessitates care in controlling the temperature, particularly on spacecraft. Nitric acid (HNO3) is commercially produced as a base material for fertilizers, dyes, and explosives. When it contains about 2% water, it is referred to as white f u m i n g nitric acid. When containing about 5 to 20% nitrogen dioxide, it is a more powerful oxidizer and is then called red f u m i n g nitric acid because of its orange to red color. So-called inhibited nitric acid (IFRNA) contains 15% NO2, 2% H20, and 1% HE All variants are hazardous on skin contact and give off toxic vapors. Inhibited red flaming acid is less corrosive than the other variants and can be used in contact with stainless steel, some a l u m i n u m alloys, Teflon, and polyethylene. 4.11.2
Fuels
Hydrocarbons, which can be derived inexpensively from petroleum derivatives, also make useful rocket fuels. Examples are aircraft jet propulsion fuel, alcohol, and kerosene. In the United States, the rocket propellant RP- 1 is frequently used. RP- 1 in contact with air self-ignites at 240~ It is compatible with aluminum, steel, copper, Teflon, and neoprene. U n s y m m e t r i c dimethyl hydrazine [UDMH, (CH3)2NNH2] is often used mixed with hydrazine (which is described later). Compared with pure hydrazine, it has the advantage of a lower freezing point (-71~ It is toxic but can be stored for prolonged times. Materials that are compatible with it are titanium, stainless steel, several a l u m i n u m alloys, and Teflon. Liquid h y d r o g e n ~ w h e n combined with fluorine (which, because of its inherent great hazard, has been used only in experimental motors) or liquid o x y g e n ~ gives the highest specific impulse that has been achieved with chemical rocket motors. Its main disadvantages are its low boiling point (20 K at 1 atm) and its low density (71 kg/m3). The low boiling temperature is the cause of inevitable boil-off during launch operations, even though the cryogenic tanks are thermally insulated as m u c h as the mass of the insulation permits. In this connection, it should be noted that hydrogen-air mixtures explode over a wide range of mixture ratios. In the presence of an ignition source such as an electric spark, at 20~ and sea level atmospheric pressure, the lower and upper limits for deflagration are 4.0% and 76%, respectively, byvolume fraction of hydrogen. The range for detonation is only slightly more narrow. Formation of ice from air moisture on the surface of the launch vehicle's tanks can be a problem because of the increased mass of the vehicle. Liquid hydrogen can be stored on spacecraft in space only for relatively short periods of time. In principle, liquid hydrogen could also be carried on space missions for m u c h longer periods and be made useful for
C H A P T E R 4 Chemical Rocket Propulsion
142
propulsion on planetary return missions. To limit the loss by boil-off, the vapor in that case would have to be recondensed. However, efficient cryogenic refrigerators of the size required for such missions still await their development. As a consequence of the low density of liquid hydrogen, tanks are large, several times the size of the tanks that contain the oxidizer. The ratio in tank size becomes larger yet when, to augment the specific impulse, the rocket motors are run hydrogen rich. A small increase in density can be obtained by cooling the hydrogen below its normal boiling point (but so that it can still be pumped) to produce hydrogen slush.
4.11.3
Monopropellants Hydrazine (Nail4) is the most commonly used monopropellant for spacecraft attitude control and station keeping. When in contact with a catalyst, it decomposes spontaneously, with a large release of thermal energy. The reaction is described by the two-stage process 3N2H4
> 4NH3 + N2
/ approx. 40%
/
4NH3
> 2N2 + 6H2
where a m m o n i a is produced partly as an intermediary species in the reaction and partly as an end product. A useful catalyst is iridium, finely dispersed on a substrate of alumina. The gradual deterioration of the catalyzer bed at the high temperature of the reaction is the main factor that limits the useful life of the thruster. The degradation is more rapid when the thrusters are repeatedly started cold. For this reason, electric heaters are used to maintain the catalyzer bed at temperatures from 100 to 300~ between pulses. A c o m m o n material for the construction of hydrazine tanks on spacecraft is titanium. Stainless steels are satisfactory for ground storage, but some steels can react over time with hydrazine and produce a solid residue that, ifnot filtered out, can plug the small propellant passages on spacecraft. A disadvantage ofhydrazine is its high freezing temperature (2~ which usually makes it necessary to control the temperature of the tanks, valves, and fluid lines on the spacecraft by electric heaters. Inadvertent freezing can lead to rupture of the tubing, followed by a catastrophic failure of the entire spacecraft. Hydrazine is a clear liquid, in appearance similar to water. It is toxic and flammable and, when in contact with skin, it will cause severe burns. It is stable up to 150~ and is compatible with titanium, stainless steel, aluminum 304 and 307, Teflon, and polyethylene. Hydrogen peroxide (H202, normally at 95% concentration) starts to decompose at 140~ with release of thermal energy. It has found use as a monopropellant for attitude control and also for the propulsion of small, simple rockets.
143
4.12 Propellant Tanks
4.12
Propellant Tanks The propellant tanks of launch vehicles are often designed as integral parts of the vehicle structure. In addition to withstanding the internal fluid pressure, they must in this case also support the vehicle's static and dynamic launch loads. These include the thrust and the bending of the structure that results from transverse accelerations, steering control, and upper atmosphere winds. Still other such loads are those produced by propellant sloshing and the loads during ground transportation to the launch site. Some large propellant tanks that have been designed have extremely thin, stainless steel skins that m u s t be pressurized at all times to prevent collapse. More common, however, are tanks constructed from a l u m i n u m alloys with skins that are reinforced by internal stringers. Oxidizer and fuel tanks on launch vehicles are arranged in t a n d e m (Figs. 4.14 and 4.15) to reduce the m a x i m u m diameter of the vehicle. This arrangement lends itself to minimizing the aerodynamic drag and effectively transmitting the thrust through the structure. Interstage Structure \
Fuel Duct (2)
Aft Skirt
/
..
\
----~,
/
Upper Stages
41[
' 9
StringersJ
I
! !
.
/
! I I ,~~
/
"~~
Main FuelTank
Main OxidizerTank
~
1)
Oxidizer Duct
Thrust Mount
Intertank Bulkhead
Figure 4.14
Typical tandem arrangement ofpropeUant tanks on a launch vehicle. (Adapted from Ref. 2.) Helium Bottles--~
Fuel D u c t - - -
Gimbal
"(Payl~ ~ ~_/i ~ __~~___ ~___'_~ Fuel Tank / 111 /4L II1 q~ .
.
.
.
Pressurization
Main FuelT a n k - / ~ Tank Support
Main OxidizerT a n k ~ / ~
Valves
nt
Booster Stage
Figure 4.15 Typical arrangement of propellant tanks and bottles for pressurization gas on an upper stage vehicle. (Adapted from Ref. 2.)
144
C H A PT E R 4
Chemical Rocket Propulsion
The required internal volume of the propellant tanks is the nominal propellant volume at normal temperature, corrected 1. For the volume of the specified propellant reserve, in some cases as much as 5%. This reserve must be sufficient to compensate for errors produced by the propellant mixing ratio control, inaccuracies resulting from temperature differences in the propellant loading at the launch site, and fluctuations in the rocket motor performance. 2. For the volume of t r a p p e d propellant, that is, propellant that is not usable because the geometrical configuration is such that the tanks cannot be emptied completely. 3. In the case of cryogenics, for the volume of the propellant that is boiled off in the time between filling the tanks and launch. 4. For the tank's ullage (a term borrowed from wine making), that is, the volume that must compensate for the differential thermal expansion of propellant and tank walls and for deflection of the walls under pressure.
4.12.1
Noncryogenic Tanks Figure 4.14 shows a typical arrangement of the fuel and oxidizer tanks on a lower stage of a launch vehicle. The large, cylindrical shape with ellipsoidal or hemispherical end caps lends itself to structural integration with the vehicle. To save weight, the fuel and oxidizer are often separated by what amounts to a single bulkhead, either forged or doubly welded to avoid all possibilities of leakage between the tanks. The tank walls are reinforced by internal stringers or else are fabricated from machined waffle grids. Internal structural reinforcements are also helpful in minimizing the sloshing of propellant induced by vehicle motion. For nonmetallic tanks, filament winding is often a preferred method to obtain the needed strength at mini m u m weight. For a fixed volume and weight, spherical tanks have the least shell stress from internal pressure. Fill, drain, pressurization, and vent openings require local strengthening of the shell to compensate for stress concentrations. Spherical tanks, however, preclude the use of their walls as load-carrying members that can transmit the thrust. Figure 4.15 illustrates spherical tanks arranged in t a n d e m on an upper stage vehicle. Support rings around the midsection connect the tanks to the load-bearing skin and stringers of the vehicle. Large tanks are fabricated from welded gores, smaller ones often by forging the two half-shells. Tank pressures typically range from 100 to 300 N/cm 2 for medium-size and smaller tanks. The pressure is maintained either by gas generators or by helium or dry nitrogen supplied from high-pressure bottles. These in turn are spherical, usually forged from titanium. The initial gas pressures in these bottles may be as high as 3500 N/cm 2 or even higher. Launch vehicles usually use turbo p u m p - d r i v e n propellant feed systems. This allows the internal tank pressure to be relatively low, ranging typically from 20 to 70 N/cm 2 (absolute) pressure. The materials used in the construction must be chemically compatible with the propellant. A brief list of such materials, together with the
4.12 Propellant Tanks
145
propellant that is in contact with them, is contained in Appendix E. Typical materials are a l u m i n u m alloys such as the 2000 and 6000 series used in the United States; steels such as the MSI series 300, A286; m o l y b d e n u m and inconel; titanium alloys such as 6 M-4V; and filament-wound composites. In addition to the fill, vent, and pressurization lines and valves associated with the tanks are such auxiliary devices as pressure relief valves, sensors for the propellant quantity remaining in the tank, pressure and temperature sensors, and fluid flow baffles. Fracture m e c h a n i c s is an important theoretical tool in the analysis of high-pressure tanks. Different safety factors are applied depending on whether operating conditions are such as to present a hazard to personnel and vital components or do not do so. Tanks are proof tested with highpressure water. This reduces the danger to personnel ifthe tank should burst. A commonly applied requirement is to prooftest to 1.25 times the m a x i m u m expected operating pressure.
4.12.2
Cryogenic Propellant Tanks The design principles described in the preceding section also apply to tanks intended to contain cryogens. Additional requirements are the choice of materials that are suitable at low temperatures and the need to thermally insulate. An important required property of construction materials is that they remain ductile at the temperature of the stored propellant. Suitable are some specialty stainless steels and nickel alloys. To prevent excessive boil-off, thermal insulation on the exterior of the tank is usually needed. For liquid hydrogen it is always needed. In this case, the temperature is so low that ambient air at the launch site will freeze on the exterior of the tank. Thermal insulation adds weight to the launch vehicle (but so does ice formed on the vehicle's surface). The need for and the degree of insulation will depend on such factors as the boiling point of the propellant at the tank's design pressure and the expected ambient conditions at the launch site. Cork and various foamed materials h a v e b e e n found to be useful materials that not only thermally insulate but also withstand the aerodynamic forces encountered during launch. A typical material is a phenolicimpregnated fiberglass honeycomb filled with thermal insulation material. The critical importance of insulating liquid hydrogen tanks, when compared with other cryogens such a s liquid oxygen, is seen from the following example. The heat of vaporization of hydrogen at 1 atm pressure is 4.61 105 J/kg, that of oxygen 2.14 105 J/kg. The densities of the liquids are 70.8 kg/m 3 and 1131 k g / m 3, respectively. Assuming a stoichiometric mixture and geometrically similar tanks, the ratio of the volumes of the hydrogen to the oxygen tank is therefore 2.01, the surface ratio the 2/3 power of this, hence 1.59. If the thermal insulation is assumed such that the heat transfer rates per unit surface area are the same for both tanks, a simple calculation shows that the time to evaporate a specified fraction of hydrogen compared with the time to evaporate the same fraction of oxygen is only 0.170.
146
C H A P T E R 4 Chemical Rocket Propulsion This rapid evaporation of hydrogen is an important consideration in launch preparations. Also, in the case of a delay, the need frequently arises to empty the vehicle tanks, only to refill them again from ground-based tanks for a later launch. Thermal insulation is also needed for intertank c o m m o n bulkheads, illustrated in Fig. 4.14. In the absence of insulation, the propellant with the lower boiling point could otherwise cause freezing of the other propellant. During launch, spacecraft are protected by the launch vehicle's shroud and are therefore not exposed to aerodynamically induced forces. Multilayer thermal blankets (described in Chap. 7) can therefore be used to protect cryogen tanks from solar thermal radiation and from the radiation generated by hot engine parts.
4.12.3
Operation of Propellant Tanks in the Weightless Condition The location of the drain port of tanks is normally such that acceleration of the vehicle will cause the fluid to move toward the port and cover it. However, during periods of zero thrust when in weightless condition, the propellants may be randomly distributed in the tanks, leaving the drain ports uncovered. If this were allowed to occur, the pressurization gas alone, or mixed with propellant, would enter the propellant feed system each time the rocket motor was started. Several methods are available to prevent this. Conceptually, the simplest is to use the reaction control system that is ordinarily used for attitude control to provide a small, brief acceleration of the vehicle along the line of thrust, thereby settling the propellant at the drain port. (Clearly, there will be an analogous problem with the reaction control propellant in its own tank; this is discussed later.) Another method makes use of the surface tension of the propellant relative to a wire screen or other mesh-type material located in the tank near the port. The objective is to confine at all times at least a portion of the propellant at the drain port. To mitigate against fluid sloshing, the screen is often placed in a small compartment, open to the propellant at one end. Especially useful for the smaller tanks on spacecraft are diaphragms. They serve to separate the propellant from the pressurization gas. An example of such a tank, used for containing attitude control propellants, is shown in Fig. 4.16. The diaphragm is designed such that when the tank is full, it lies against the tank wall on the gas side. As propellant is being expelled, the diaphragm, in a rolling motion in contact with the wall, moves toward the drain port. Diaphragms are frequently made from elastomers. Some propellants, particularly nitrogen tetroxide, however, tend to degrade them over long periods of exposure. For this reason metallic diaphragms are often preferred. To prevent corrosion and to facilitate welding of the edge of the diaphragm directly to the tank, an alloy similar to the one used for the tank shell is often chosen. To minimize the pressure difference between gas and propellant and the stress induced in the diaphragm by its bending, metallic diaphragms as thin as 0.25 m m have been used.
4.13 Propellant Feed Systems of Launch Vehicles
Pressurization
j
147
Inlet
Propellant~ ~ Outlet
DiaphragmPositions
Figure 4.16 Tank with diaphragm for expulsion of propellant in a weightless condition. In contrast to diaphragms, elastomeric bladders contain the entire propellant. They are compressed by the gas that is fed to the space between tank wall and bladder.
4.13
Propellant Feed Systems of Launch Vehicles In the early days of experiments with liquid-propellant rockets, a very simple system that relied exclusively on gravity and acceleration was used to force the propellants into the rocket motor. The American rocket pioneer Robert H. Goddard, who in 1926 flew the first successful liquid-propellant (liquid oxygen and gasoline) vehicles, demonstrated rockets of this type. Although simple in concept, this system, however, cannot provide the large propellant flow rates that are required by today's heavy launch vehicles. To obtain these, turbo p u m p s are needed.
4.13.1
Feed System Cycles Shown schematically in Fig. 4.17 are several variations of propellant feed systems for launch and upper stage vehicles. In all of these versions, the fuel and oxidizer are brought by p u m p s from the relatively low pressure in the tanks to the much higher pressure required for injection into the thrust chamber. This high pressure is also needed to produce a fine spray as the propellants enter the chamber, thereby promoting rapid mixing of the propellants and ensuring stable combustion. The pumps are driven by high-speed, supersonic gas turbines. A single turbine and the two pumps are sometimes arranged on a single shaft. But more often a gear train between turbine and p u m p s is preferred. The aim is to reduce the high rotational speed of compact, efficient turbines to the lower speed that is tolerated by fluid pumps.
148
C H A P T E R 4 Chemical Rocket Propulsion
Figure 4.17 (a) Gas generator system; (b) thrustchamber bleed-off; (c) dual combustion system; (d) cryogenic fuel expander cycle. F, fuel; O, oxidizer; FP, fuel pump; OP, oxygen pump; T, turbine; GG, gas generator. In version (a) shown in the figure, relatively small flows of fuel and oxidizer are diverted to a gas generator, which produces the gas needed to drive the gas turbine. The exhaust from the turbine is d u m p e d overboard in a direction to add to the thrust. It can also be used for injection into the base region, thereby improving the performance of aerospike nozzles (Fig. 4.10). The main flow of the fuel, before being injected into the thrust chamber, flows through the nozzle tubular structure (Fig. 4.2), thereby cooling it. Other arrangements are possible, such as separate turbines for each pump, with gas supplied from a gas generator either in series or in parallel. In place of the fuel, the oxidizer has sometimes been used as the coolant. The schema shown in (b) is similar to (a), except that in place of the gas generator a hot-gas bleed-off from the thrust chamber drives the turbine. In (c) is shown an arrangement known as a dual or staged combustion system. Here, the fuel, after cooling the thrust chamber, is combined in the gas generator with a m u c h smaller flow of oxidizer. The resulting fuel-rich combustion gas is designed to be at a temperature compatible with the turbine's
4.13 Propellant Feed Systems of Launch Vehicles
149
temperature limitation. This gas drives the turbine and then is led to the thrust chamber, where it combines with the balance of the oxidizer. In (d) is shown the cryogenic fuel e x p a n d e r cycle. For hydrogen-oxygen propulsion, this cycle has become one of the preferred arrangements. Here the gaseous hydrogen that exits from the nozzle wall tubes expands first in the turbine and then enters the thrust chamber, where it combines with the oxygen. C o m m o n to all systems that are illustrated in Fig. 4.17 is so-called regenerative cooling of the nozzle ("regenerative" because the heat added to the propellant is not lost but added to the subsequent combustion). Typically, the propellant is routed through the nozzle tubes from the nozzle exit plane to the thrust chamber injection plane ("single-pass cooling") or else startsat the injection plane and returns to it in adjacent parallel tubes ("double-pass cooling"). The tubes, typically of a nickel alloy, are deformed from their circular cross section to fit the nozzle contour but also to obtain m a x i m u m flow velocity, hence m a x i m u m heat transfer rate, near the throat, where the heat transfer from the combustion gas tends to be highest. A consideration that sometimes decides the relative advantages i n regenerative cooling ofusing either the fuel or the oxidizer is potential pinholesize leaks from the coolant tubes into the thrust chamber. If the combustion gas is fuel rich, as is usually the case for hydrogen, a small leak of fuel into the nozzle interior may be harmless, since the leaking fuel will tend to cool the damaged spot rather than cause a burn-through, as would be likely to occur if the oxidizer had been employed as the coolant. To supplement regenerative cooling, film cooling of the thrust chamber interior surface in the combustion zone is often used. For this purpose, a portion of the fuel is injected in a direction to wet the wall. The resulting fluid film, before its evaporation, can in some cases cover the thrust chamber surface nearly up to the nozzle throat. The fuel flow used in this manner, however, must be kept relatively small so as to avoid incomplete combustion. Film cooling is also often provided to protect the walls of gas generators. The turbines that drive the propellant p u m p s are similar to aircraft jet engine turbines. The reader is assumed to be familiar with their basic design or may wish to consult a standard reference (e.g., [9]). Figure 4.18 illustrates a typical design of a two-stage turbine for launch vehicles. To achieve a compact and efficient design of a turbo pump, the turbine tip speed must be considerably higher than the tip speed of the pumps. This condition is often met by introducing gears between turbine and pumps. A reduction of the shaft speed from turbine to p u m p by a factor of about 3 is quite common. The useful life of turbo p u m p s on expendable launch vehicles is counted in minutes. The gear loading (torque and speed) can therefore be chosen to be extremely high to ensure a compact and lightweight design. To avoid almost immediate destruction by overheating, a large flow of lubricant is p u m p e d through the gear case. An example of a turbo p u m p with a twostage gas turbine and a two-stage reducing gear train driving the fuel and oxidizer p u m p s on a c o m m o n shaft is shown in Fig. 4.19. In designs that omit gears, the turbine diameter is chosen to be substantially larger than the p u m p diameter. This allows a relatively high tip speed of the turbine in combination with a lower tip speed of the pumps, as needed to avoid excessive cavitation.
1S0
C H A P T E R 4 Chemical Rocket Propulsion
Figure 4.18
Two-stage gas turbine for driving propellant pumps. (Adapted from
Ref. 2.)
Figure 4.19 Geared turbo pump" oxidizer and fuel pumps on a common shaft driven by a two-stage turbine and two-state reducing gear. From Ref. 2, Huzel, D. K., et al., "Modern Engineering for the Design of Liquid Propellant Rocket Engines.', Courtesy of Rocketdyne Division of Rockwell International. Copyright 9 1992, AIAA-- reprinted with permission.
4.13 Propellant Feed Systems of Launch Vehicles
4.13.2
151
Propellant Pumps Because of the occurrence of cavitation, which depends on the vapor pressures of the propellants, p u m p s on launch vehicles deserve a somewhat more detailed discussion here. Cavitation refers to the formation of vapor and gas bubbles in the lowpressure regions of the flow of a liquid. Bubble formation can also be excited acoustically, or simply by boiling; what is characteristic, however, of the p h e n o m e n o n of cavitation is the rapid, and often destructive, collapse of the bubbles as they are transported bythe fluid from a low- to a high-pressure region. In the collapse, the enclosed vapor and gas revert to the liquid phase or are absorbed by the fluid. The dynamic forces resulting from the collapse can be large enough to pit and destroy rapidly metallic surfaces in the vicinity. In addition to its destructive effect, cavitation is deleterious to the mechanical efficiency of p u m p s and can induce a nonsteady flow resulting in large, possibly destructive, pressure fluctuations in the propellant ducts. An example of such a nonsteady flow is "rotating cavitation," which has some similarity to the rotating stall of compressors. If a vapor bubble were in a state of equilibrium, the pressure in its interior would be the vapor pressure of the surrounding liquid, modified by the surface tension at the interface. For this reason, a rough estimate for the occurrence of cavitation can be obtained from the relation p < pv(T)
(4.48)
where p is the local pressure of the fluid and pv(T) its vapor pressure at the operating temperature. Large departures from this simple criterion can occur in practice [10]. In very pure liquids, produced under laboratory conditions, the fluid can sustain several atmospheres of tension. Therefore the fluid pressure can be substantially below the vapor pressure without bubble formation. On the other hand, in propellants and other engineering fluids, such large tensions cannot occur because sufficient numbers of nucleation sites will be present either in the fluid or on containing surfaces. Nuclei that are suspended in the fluid can be either small solid impurities or microscopic bubbles formed from gases in the fluid. Gas bubbles that are already present ahead of a p u m p can grow in size in the pump's low-pressure regions, even though the fluid pressure is well above the vapor pressure. This phenomenon has been called "pseudocavitation." The vapor pressures of rocket propellants are well known, but the effect of particles and dissolved gases on the critical pressure must be determined case by case in the laboratory. Cavitation, if present, will tend to occur near the p u m p inlet and on the suction side of the p u m p blades, where the pressure is lowest. As a consequence of the curvature in the'meridional plane of the flow passage of centrifugal p u m p s (Fig. 4.20), the most prevalent location in these p u m p s for cavitation to occur will be near the tip of the blades at the inlet. Propellant p u m p s on launch vehicles can be of the axial flow or centrifugal type. The latter are often preferred because they allow a highly compact arrangement and produce a large pressure ratio in a single stage.
152
C H A P T E R 4 Chemical Rocket Propulsion
Figure 4.20 Typical propellant pump. The centrifugal p u m p illustrated in Fig. 4.20 is quite typical. The fluid enters the p u m p in the direction parallel to the axis of rotation. It then turns toward an approximately radial direction, thereby increasing the pressure as a result of the centrifugal force. The rotating blades that are attached at their hub to the rotor have an approximately spiral shape. One sometimes distinguishes between the "inducer" and the "impeller" sections. In the former, the flow is nearly axial and the blades are designed to produce only a modest pressure rise. Most ofthis rise then occurs in the impeller section and in the diffuser, where stationary blades are arranged to turn the flow into the circumferential direction of the volute. The latter is given a cross-sectional area that increases approximately linearly with the azimuth. It discharges the fluid into the high-pressure duct. The p u m p shown in the illustration has an impeller with a "shroud." Its purpose is to minimize fluid leakage at the blade tips. An approximate, theoretical method, sufficient for a first estimate of the pressure rise, is based on a model in which the blades and their action on the fluid are replaced by a circumferentially uniform body force (infinite n u m b e r of blades model). One then considers a streamline, say S, that is intermediate between the hub and the blade tips and spirals in conformity with the pitch angle of the blades. If u and p are the fluid velocity and pressure, 0 the density, and 9 the potential of a conservative force field that acts on the fluid, one form of Bernoulli's equation for the steady flow of an inviscous, incompressible
4.13 Propellant Feed Systems of Launch Vehicles
153
~.=~. arge
~5
-" " _ &l
~
Discharge
~
Inlet ~
0
Inlet (a)
(b)
Figure 4.21 Schematic of centrifugal pump. (a) impeller; (b) velocity diagram. (1) inlet; (2) transition from impeller to stationary blades; (3) volute. fluid is 1 ua + p § 9 2 0
constant along streamlines
(4.49)
The potential of the centrifugal force associated with the rotating parts of the p u m p is -~oar a/2, where co is the angular velocity and r the distance from the axis. Applied to the axisymmetric surface defined by S, this b e c o m e s 1
2_1 /~'
2
0
a~2f~__l
2
24 /~2
2
O
~2f~
(4.50a)
2
where g~ is the magnitude of the velocity in the corotating frame of reference and the subscripts ( )1 and ( )2 designate the inlet to the inducer and the exit from the impeller, respectively. (It m a y be noted here that neither the Coriolis force nor the body force representing the action of the blades on the fluid contributes to the energy increment along the streamlines, hence to Bernoulli's equation, because both forces are perpendicular to the streamlines.) The various velocity vectors at an arbitrary point on S, on the rotating parts, are illustrated in Fig. 4.21a. Here ( )~ and ( )0 designate the meridional and circumferential components, respectively. The velocity c o m p o n e n t s and tb relative to the stationary and rotating parts, respectively, of the p u m p are related by g;0 + a# = v0 For the stationary blades @ = 0. From (4.49),
~
+ P2 1P3 = ~v~ -tQ e
(4.50b)
154
C H A P T E R 4 Chemical Rocket Propulsion where ( )3 designates the condition at the trailing edges of the stationary blades. As a further approximation, this latter condition is taken to be the same as the condition at the discharge end of the volute. Figure 4.21b illustrates the velocity vectors at the three stations (1), (2), and (3) that are being distinguished in the present approximation. Conventionally, the angles fl and y, which are measured in the plane tangential to the axisymmetric stream surface, are called the "pitch angles" of the flow at [he leading and trailing blade edges, respectively. In an actual pump, these angles differ somewhat from the true blade angles because of the effect of the finite n u m b e r of blades on the cascade's inflow and discharge angles. The pitch angles of the flow leaving the impeller must be compatible with the pitch angle of the flow entering the stationary blades. As is readily seen from the vector diagrams in the figure, this condition, for the streamline S, is cotan Y1 -[- cotan f12 = _wf2
V#,2
(4.51 )
Finally, if b is the width of the flow channel in the meridional plane and r~ the mass flow rate, conservation of mass requires that 0~ - @~ -
m 2zr f bo
(4.52)
Substitution of (4.52) into (4.50a) results in the expression
1(~2_ ~1)_ 1 (
-2 ) l a j 2 sin 2 Yl
for the pressure increase from the inlet to the trailing edge of the impeller. Similarly, the pressure increase from the leading edge ofthe stationaryblades to the volute discharge is given by
_l(p3 _ p2) _ 1 ( 02~2. 0
2
sin2 F]
02 --09f20#,2 cotan Yl -~- ~lo92f2
Adding the two equations and introducing the additional assumption that the design of the p u m p is such that the discharge velocity ~3 equals the inlet velocity ~1, one obtains the final result for the ideal pressure rise that can be obtained from the p u m p
1 Q
--(P3 -- Pl) -- 092/:2--0)r20#,2 cotan Yl
(4.53)
High discharge pressures of the propellant p u m p s are necessary for injecting the propellants into the thrust chamber with a high pressure difference so as to form a spray of small droplets conducive to thorough mixing of the propellants. Also, the thrust chamber itself needs to be at a high pressure so as to avoid overexpansion. On the other hand, the inlet pressures must be chosen by a compromise between the need to mitigate against cavitation and the need to reduce the weight of the tank structure and therefore the tank pressure.
4.14 Thrust Chambers of Liquid-Propellant Motors
155
Computational fluid mechanics provides the theoretical means for more accurate calculations of the flow. In particular, the correct geometry of the blade cascade can be taken into account. Results of such calculations reveal strong secondary flows in the space between the blades. These secondary flows are similar to those observed in bent pipes or elbows. Finite Reynolds n u m b e r effects also can be taken into account. A requirement c o m m o n to all propellant feed systems is the need to start and shut down the rocket motor safely with the least loss of propellant. In systems with gas generators, these are activated before engine start so as to bring the turbo p u m p s up to speed. During start-up as well as during shutdown, the propellant p u m p s must operate at off-design conditions. It is important then to avoid conditions that result in massive stall with attendant large pressure fluctuations and vibrations. Some designs therefore incorporate bypass lines and valves that during start and shutdown route parts of the propellant from the p u m p discharge back to the p u m p inlet. Some engine designs allow not only multiple starts and shutdowns but also variations in thrust by throttling. Here, too, avoidance of stall of the p u m p s is an important consideration. The needed control is facilitated if the fuel and oxidizer p u m p s are driven by separate and separately controlled turbines.
4.14
Thrust Chambers of Liquid-Propellant Motors An important element of the thrust chamber is the injector. In the design illustrated in Fig. 4.22, the oxidizer enters the dome above the injector plate, then flows through a large n u m b e r of small holes in the meridional planes of the injector plate before being injected at high pressure into the thrust chamber. Similarly, the fuel is admitted from an annular manifold through radial holes in the plate. The injection orifices (ports), separate for oxidizer and fuel, cover the entire circular area of the lower side of the injector in the m a n n e r of a shower head.
~ . -. _•___••_ oc, "-,~
Injection //-Plate
oOC 000
I
Oxidizer and / Fuel Sprays
Figure 4.22
Thrust chamber injection plate.
156
C H A PT E R 4
Chemical Rocket Propulsion
Breaking up the propellant streams into small droplets is aided by the high pressure drop in the orifices. It can be further improved by arranging the angles of the propellant passages in the injector such that the oxidizer jets, and similarly the fuel jets, impinge on their neighbors. Alternatively, the oxidizer jets are made to impinge on the fuel jets and vice versa. In hydrogenoxygen rocket motors the injection of the gaseous hydrogen, at sonic velocity in the ports, is from openings that surround coaxially the injection ports of the liquid oxygen. A special problem arises when throttling of the engine over a large range of thrust is required, as is the case in lunar or planetary descent engines. To ensure that the pressure drop at the ports is sufficient to atomize the liquids, movable pintles have been used successfully. The droplets quickly vaporize by heat transfer from the surrounding combustion products. The gas-phase reaction that follows is aided by the rapid diffusion of the reactants and intermediate combustion products. Even for propellants that self-ignite on contact, an igniter needs to be provided to ensure rapid ignition at engine start. Otherwise, because the pressure in the thrust chamber at the start is low and the reaction may be incomplete, a detonation of unburned propellants in the thrust chamber and beyond could occur. The m i n i m u m needed length of the combustion zone is dictated by the residence time of the propellants that is needed for complete evaporation, mixing, and combustion. Depending on the pressure and type of propellants, this time may vary from 2 to 50 ms. Most of it is required for evaporation and mixing, rather than for the combustion kinetics. Computer models that have attempted to describe in some detail the droplet formation, evaporation, and combustion reaction are useful for extrapolating from known thrust chamber designs to modified designs but otherwise have had only limited success. The design of thrust chambers requires extensive testing. An important part is the verification that no major, possibly destructive, combustion instabilities are present. Several types of such instabilities have been observed. They have been classified into intrinsic acoustic, injection-coupled acoustic, and lowfrequency oscillations. Intrinsic acoustic oscillations occur at the natural acoustic frequencies characteristic of the chamber geometry. They are therefore associated with sound and shock waves that traverse the chamber at sound speed or higher. Both standing and propagating waves, either longitudinal "organ pipe," radial, or circumferential oscillations have been observed. Potentially the most destructive ones are the circumferential oscillations. Whereas their frequencies can be calculated approximately from standard acoustic theory, much more difficult to predict are the amplitudes, as these can depend on the processes of propellant atomization, evaporation, mixing, and combustion. Injection-coupled acoustic oscillations depend on the injection velocity, hence on the pressure difference across the injection orifices. Gas pressure fluctuations in the thrust chamber influence the injection pressure difference, hence couple to the propellant injection rate, which in turn
4.14 Thrust Chambers of Liquid-Propellant Motors
157
induces fluctuations of the rate of propellant burned and therefore of the gas pressure. The stability margins can be improved by injecting at high pressure, at the cost, however, of increased propellant pump power. Modes of instability in which the combustion chamber pressure is more or less spatially uniform, but fluctuates in time, are referred to as lowfrequency oscillations or chugging. They are caused by the coupling of the chamber pressure with the pressure in the propellant feed system. An important form of low-frequency oscillations, called pogo (by analogy with the jumping stick toy), involves, in addition to the thrust chamber and the propellant feed system, the structural modes ofvibration, particularly the longitudinal ones, of the entire vehicle. The pogo instability is more amenable to analysis then the high-frequency oscillations that depend intimately on the processes of combustion. It is briefly described in the next section. Avoidance of dangerous, high-frequency combustion instabilities requires a largely empirical approach. Modifications to the injection pattern, providing baffles, or providing acoustic cavities in the thrust chamber, have been successfully employed. Small, random pressure fluctuations as high as 5% of the average pressure must be expected even with optimized injection and thrust chamber designs. They are responsible for the very large acoustic output of rocket motors. The resulting sound and shock waves propagate in the atmosphere externally to the launch vehicle but also propagate through the vehicle's structure and cause vibrations and shocks. If unprotected, electronic and other sensitive components either in the vehicle or stationary at the launch site can be damaged. The sound waves that emanate from the engines and that propagate through the atmosphere at the launch site can even affect components on spacecraft on top of the vehicle, even though they may be partially protected by the payload shroud. The sound pressure levels can reach 140 dB or more, with power spectral densities peaking between about 200 and 2000 Hz. For vibrations that propagate directly along the vehicle structure, the natural damping by the materials causes the spectra of random vibrations to shift toward lower frequencies, with power spectral densities primarily between 50 and 1000 Hz, depending on the location on the launch vehicle. It is important, for instance in the design of shelves that support electronic hardware on spacecraft, to guard against resonances in this frequency band. Mention has already been made of the need for cooling the thrust chamber. The rate of heat transfer by convection from the combustion gas to the thrust chamber walls tends to be highest at the nozzle throat or slightly upstream of it. At this location the combination of the still high gas density and of the already large Mach number (Fig. 4.6) causes a large heat transfer rate. In comparison with convective heat transfer, radiative transfer in liquid propellant motors is smaller. Nevertheless, depending on the type of propellants and size of the motor, radiative transfer can still be significant. At least in very large motors the gas can become partially opaque and in the extreme case would therefore radiate as a blackbody. Two-atom molecules such as hydrogen, oxygen, or nitrogen do not have emission or absorption bands in the wavelength regions that are pertinent for rocket propulsion. On
158
C H A PT E R 4
ChemicalRocketPropulsion
the other hand, heteropolar gases, such as water vapor, carbon monoxide and dioxide, the oxides of nitrogen, and ammonia, have strong bands. Even so, their contribution to the total gas-side heat transfer in liquid propellant thrust chambers rarely exceeds 10%, and less in small motors, where the gas is essentially transparent. The convective heat transfer rate, hc = ~//A T (~/- heat transferred per unit time and unit area; A T = gas-to-wall temperature difference) can be estimated from the slightly modified form of the equation for turbulent flOWS, Nu = C Re ~ Pr T M
( 4 .5 4 )
where Re = oud/lz is the Reynolds n u m b e r referred to the diameter, d, Pr = ~ c p / k the Prandtl number, and Nu = hcd/ k the Nusselt number (u = velocity; 0 = density; lz = viscosity; k = gas thermal conductivity; cp = specific heat at constant pressure). The nondimensional constant C depends on geometrical factors. For straight tubes, C = 0.024. 0, lz, k, and cpare obtained from the sum ofthe mass-weighted amounts of the constituents in the gas mixture. An approximate value for the Prandtl n u m b e r can also be calculated from the formula Pr =
4y 9y - 5
(F = ratio of the specific heats of the gas mixture) obtained from gas kinetic theory. For several reasons, (4.54) may underestimate the true heat transfer rate. The addition from radiation has already been mentioned. Recombination at the wall and in the boundary layer ofdissociated species will also increase the heat transfer. Semiempirical formulas that are modifications of (4.54) and are based on rocket motor tests have been proposed by Bartz and others.
4.15
Pogo Instability and Prevention A n u m b e r of launch vehicles have exhibited an instability referred to as p o g o . It shows itself as a low-frequency vibration of the entire vehicle, often along its longitudinal axis. The vibration may start some time after launch, grow, and then decay again. Peak accelerations as high as 300 m / s 2 have been observed. The pogo instability and means to suppress it have been studied by Rubin, 1970; Oppenheim and Rubin, 1993 [11]; and others. The instability arises from the interaction of the vehicle's structural modes of vibration with thrust oscillations. These in turn involve the thrust chamber and the propellant feed system. The analysis is based on the mathematical description of the closed-loop process that includes all of these subsystems. Simultaneously with the occurrence ofpogo, there will also be parametric changes ofthe vehicle properties, particularly due to propellant depletion and the consequent change in vehicle mass. These, however, are slow compared with the pogo frequencies. The latter can therefore be analyzed as being superposed on a pseudostationary process.
159
4.15 Pogo Instability and Prevention
9 Nodes
LOX
Bellows (~
Pumps Accumulators
~
Fuel
Junctions Fuel ~
Sustainer Motor
/ Booster
Booster
Figure 4.23 Nodes for pogo oscillation analysis. (Adapted from Oppenheim and Rubin [11].) The analysis starts by defining as discrete nodes all the components that are likely to play a role (Fig. 4.23). These include the thrust chambers, propellant pumps, ducts, their junctions, bellows, tanks and their outlets, hydraulic accumulators, and the principal vehicle structural elements. Each node will receive inputs from, and provide inputs to, some other nodes. It is important to include the elastic compliance of elements such as tank walls and bellows. More difficult is the proper representation of the contributions from thrust chambers and pumps. These are largely empirical. For instance, the representation of the pumps must include the variability of the flow rate-dependent cavitation in the inducer section and the performance changes caused by variations in the blades' angles of attack. The importance of including among the variables the degree of cavitation has long been recognized [11]. The properties of each node are described by first- and second-order, linear differential equations in the state variables, for which convenient choices are the pressures, flow rates, and structural displacements. This coupled system of equations can be analyzed by classical methods that yield the complex eigenfrequencies and eigenmodes of the system. It has been demonstrated that the pogo instability can be suppressed by introducing hydraulic accumulators ("pogo suppressors") into the propellant feed system. The accumulators are therefore an essential element in the stability analysis.
160
C H A P T E R 4 Chemical Rocket Propulsion
4.16
Thrust Vector Control There are several ways to alter the direction of the thrust for steering the rocket. These methods are jointly referred to as t h r u s t vector control. Most frequently used is a method based on gimballed t h r u s t chambers (Fig. 4.24). The gimbals are on two perpendicular axes that allow independent rotations of the thrust chamber. The maximum rotations allowed by the gimbals are typically angles of about • ~. The rotations, and therefore the vehicle steering, are imposed by actuators, which, for large motors, are usually hydraulic but can also be pneumatic or electric. To accommodate the motion ofthe thrust chamber, the propellant ducts must have the necessary flexibility. This is accomplished by a series of bellows, usually made of stainless steel. Bellows are a critical element of gimballed chambers because their skin must be thin enough to offer only limited
Gimbal
Gimba' I I
i
Axis~ .
Bellows
/ /
Nozzle
I i
Figure 4.24 Gimbal axes and propellantline bellows of the LEM engine, NASA. From Ref. 2, Huzel, D. K., et al., "Modern Engineering for the Design of Liquid Propellant Rocket Engines." Courtesy of Rocketdyne Division of Rockwell International. Copyright 9 1992, AIAA~ reprinted with permission.
4.16 Thrust Vector Control
161
resistance to flexing, yet strong enough to withstand the pressure. It is evident that if the motion is limited to, say, +7 ~ in both directions, the bellows must allow thrust c h a m b e r motions of this order of magnitude. The motions of the hydraulic or p n e u m a t i c actuators are controlled by servo-valves, which in turn are controlled by c o m m a n d s from the vehicle's guidance system. Displacement and rate transducers are attached to the actuators. They provide error signals that are the difference between the actual and desired positions and rates. In the control computer, closed-loop control is applied that makes use of the rate signals to ensure stability. In place of rate transducers, digital differentiation of the position signals is also being used. The figure shows a configuration in which the gimbal axes are placed approximately in the plane of the nozzle throat. Alternatively, they are often placed back of the injection plate and d o m e (Fig. 4.2). Figure 4.25 shows the gimbal actuator m o u n t s and articulating propellant ducts in the gimbal plane of the U.S. Space Shuttle main engines. Another m e a n s of thrust vector control is by lateral injection into the nozzle. Three such systems are schematically illustrated in Fig. 4.26. In all cases, gas or liquid is injected from the nozzle wall downstream ofthe throat. A shock front is formed that deflects the main gas stream. The resulting asymmetry of the flow at the nozzle exit plane causes a torque about the center of mass of the vehicle, sufficient to steer it by closed-loop control as c o m m a n d e d by the guidance computer. Four injection ports, each with its servo-valve and spaced 90 ~ apart, are needed. No more than two adjacent injection ports operate at the same time.
Figure 4.25 Articulating ducts in the gimbal plane of the Space Shuttle main engines, NASA. From Ref. 2, Huzel, D. K., et al., "Modern Engineering for the Design of Liquid Propellant Rocket Engines." Courtesy of Rocketdyne Division of Rockwell International. Copyright 9 1992, AIAAm reprinted with permission.
162
C HA PT ER 4
L, S-"
Chemical Rocket Propulsion
is_ ,
JfL G
S-" |
(a)
(b)
(c)
Figure 4.26 Lateral injection systems for thrust vector control: (a) gas chamber tap-off system; (b) bipropellant gas-generator system; (c) liquid system. S, shock front; G, hot gas duct; O, oxidizer; F, fuel duct; L, liquid duct; GV, gas valve; GG, gas generator; LV, liquid injection valve. In (a), the injected gas is tapped off from the high-pressure region in the combustion chamber. At maximum steering torque, the flow rate of the tap-off gas is typically 1.5 to 2.5% of the primary rocket gas flow. In (b), oxidizer and fuel are combined in gas generators, one for each injection port. A less frequently used scheme in which an inert fluid is injected is illustrated in (c). Compared with gimballed thrust chambers, a disadvantage of lateral injection is the need to provide ports in a thermally stressed part of the nozzle. This will increase the complexity of providing coolant passages for regenerative cooling. The disadvantage is offset by the simplicity of a fixed mounting of the thrust chamber with rigid propellant ducts.
4.17
Engine Control and Operations The principal function ofthe engine control is to ensure the correct flowrates of the propellants. In particular, the flow rates must be such as to result in the optimum mixture ratio for the flight condition at the time ("mixture ratio control"). Also, the flow rates need to be controlled so that toward the end of the firing with the tanks nearly empty, the correct ratio of the remaining fuel and oxidizer is maintained in the tanks ("propellant utilization control"). Many rocket motors are designed for a nominally constant thrust, but others allow the thrust to be varied in response to c o m m a n d s from the flight computer ("thrust control"). The principal components needed for the latter type of engine are indicated in the schematic, Fig. 4.27. Omitted for clarity are various secondary controls such as the tank pressurization controls, safety controls, and the controls needed for start-up and shutdown. The propellant utilization control is based on inputs from transducers that measure the fuel and oxidizer masses in the tanks. Acoustic sensors, capacitance probes, or differential pressure sensors are being used for this purpose. The propellant utilization control is important because there are a n u m b e r of error sources that can affect the propellant masses actually
163
4.17 Engine Control and Operations (~ Summing Function [~> Amplifier
Chamber Pressure Reference
IL~ I
Fuel Mass - - ~ FlowFUeIMeterI - I Valve Fuel t Transducer I ~ I . I~__ Flight ~--I Thrust. I Computer PU Control MR Control q Control I Reference / Reference
Reference
l Oxidizer Tank
Oxidizer Mass ~ Transducer
erence
Oxidizer I Flow Meter I
I
Oxidizer Vernier ctuator I Oxidizer ~-- Valve
I
I I I
T
J
Figure 4.27 Principal components of a typical liquid-propellant control with provision for variable thrust. MR, mixture ratio; PU, propellant utilization. present in the tanks. These include, for instance, losses of cryogenic propellants from boil-off. Other errors can be present in the initial loading as measured by load cells at the launch site. Without close control, the fuel and oxidizer masses remaining in the tanks toward the end of the firing may significantly deviate from the intended mixture ratio. The output of the closed-loop propellant utilization control is one of the inputs to the mixture ratio control. Additional inputs are those from the flow meters for fuel and oxidizer. The control is designed to maintain a nearo p t i m u m mixture ratio, independent of variations of propellant temperature, density, tank pressure, and vehicle acceleration, although modified, if needed, to satisfy the propellant utilization requirement. In particular, the effect of vehicle acceleration on the mixing ratio can be significant, mostly for the propellant in the forward tank because of the long fluid column connecting it to the engine. Whereas most engine controls are designed to maintain a near-constant mixture ratio, there can be some advantage in purposely modifying the ratio in flight. The heavier propellant is then used at the early parts of the flight at a rate slightly faster than normal and is used slightly less toward the end. The advantage is a small reduction of the gravity loss. The output from the mixture ratio control is sent to the vernier actuator, usually on the oxidizer side. It provides the fine adjustment of the mixture ratio. Engines that allow throttling also require a t h r u s t control. A suitable, although indirect, m e a s u r e m e n t of the thrust in flight can be obtained from the combustion chamber pressure. A closed-loop control then governs the positions of the main fuel and oxidizer control valves. Other control schemes are also being used. In all cases is it important to consider the dynamic properties of the components that are being regulated. They can often be characterized by their time delay, for instance, the closing
164
C H A P T E R 4 Chemical Rocket Propulsion time of valves. Also significant can be th e rotational inertia of the turbo p u m p s and the delay in the pressure buildup of gas generators. A frequently used type of control is represented by the "proportionalintegral-differential" control law Y - Yo - kle + -k21tc e dt + 751 ,]0
de
k3ZD~ dt
(4.55)
with y as the control variable. Here, e is the error term; kl, k2, ka are gains; ri is the integration time; ro is the differentiation time; and Y0 is the desired value of the control variable. The addition of the integral term in this equation eliminates the offset inherent in simple proportional controls (but may also cause overshoots). The addition ofthe differential term provides a faster transient response to rapidly varying conditions. Typical start-up and thrust cutoff operations are indicated in Fig. 4.28. At engine start, depending on the type of propellant and the engine cooling method, either the fuel or the oxidizer flow may lead. Safety considerations are essential elements ofengine start and engine shutdown. Usually, the propellant cutoff at engine s h u t d o w n is such that the condition in the thrust chamber is made fuel rich. Precise thrust cutoff at the time c o m m a n d e d by
Figure 4.28
Typical sequence of liquidpropellant motor start and cutoff.
4.18 Liquid-Propellant Motors and Thrusters on Spacecraft
165
the flight c o m p u t e r is particularly important for u p p e r stage motors w h e n the exact instant of shutdown dictates the trajectory after final orbit insertion. Omitted from the diagram are a n u m b e r of secondary operations, such as purging and ~ for cryogenic propellants ~ chill-down that must precede engine start. Launch vehicles and upper stage vehicles are often equipped with their own flight computers. It is also possible, and often advantageous, to combine all computers into a single one located in the u p p e r m o s t stage or e v e n ~ in case a single spacecraft is being launched ~ in the spacecraft itself.
4.18
Liquid-Propellant Motors and Thrusters on Spacecraft Liquid-propellant rocket motors on spacecraft are of two somewhat different types, mostly distinguished by their thrust. In the first category are the spacecraft m a i n propulsion motors. Their applications include final orbit insertion (unless this task is taken over bythe final upper stage vehicle), retrofiring prior to a lunar or planetary landing or earth return, and ascent from the surface of a planet or moon. In a second category are smaller motors, often referred to as t h r u s t e r s or reaction control motors. Typical applications are attitude (orientation in space) control of spacecraft and station keeping (keeping the same station on the geosynchronous orbit by correcting for perturbing forces). Other applications are the spin-up and spin-down of spin-stabilized spacecraft and midcourse corrections on deep-space missions. The b o u n d a r y between these two categories is fluid. A useful distinction that affects the level ofthrust and therefore the size ofthe engine can be m a d e between maneuvers in a gravitational field and those in a near-weightless environment. In the first case, the time allowed for firing is relatively short, usually counted in tens of seconds or a few minutes. To achieve a desired total impulse, the motor must therefore be relatively large. In the second case, such as in orbit corrections and in several applications to deep-space missions, the period of thrust can be quite long because there will be no gravity loss. Thrust periods of the order of an hour or more are possible. Hence a m u c h smaller motor will be adequate to achieve the needed total impulse. Except for size, liquid-propellant spacecraft motors are conceptually similar, yet simpler than launch vehicle motors. The propellant flow rates are lower and the tanks are smaller and can be designed for higher pressure. In most cases this obviates the need for turbo pumps. Instead, the simpler system can be used that involves pressurized gas expulsion ofthe propellants from the tanks. As an example of a spacecraft m a i n propulsion motor, one of the NASA Viking engines is shown in Fig. 4.29. This engine was one of three identical terminal descent engines, spaced equally around the circumference of the craft, and activated shortly before landing on the Mars surface. The propellant load was 85 kg hydrazine, contained in two titanium spheres pressurized with dry nitrogen at an initial pressure of 360 N / cm 2. The deceleration of the
166
C H APT E R 4 Chemical Rocket Propulsion
Figure 4.29 Terminal-descent motor for Viking Mission. (From Holmberg, Faust, and Holt, NASA Reference Publication 1027.)
craft was sufficient for the propellant to cover at all times the tank exit ports. Therefore no diaphragms were needed. Thrust variation was provided by a motor-driven throttle valve that controlled the hydrazine flow. The maxi m u m thrust was 2650 N. To avoid generating a dust cloud at the landing site, the exhaust of each engine was divided a m o n g six nozzles. By way of contrast, Fig. 4.30 illustrates a reaction control m o t o r (thruster). Although monopropellant motors are often preferred for their greater simplicity, bipropellant motors, such as the one shown, have a larger specific impulse. The propellant flow rates are controlled by solenoid valves. Cooling is partially by ablation of the carbon-phenolic nozzle and its graphite insert and partially by radiation. Thrust vector control is also greatly simplified, because the desired result can be achieved by changing the attitude of the entire spacecraft either before or even during the firing of a main motor. Small t h r u s t e r s are useful for attitude control and station keeping. To allowrotation ofthe spacecraft about three axes and, independently, acceleration along these axes, a m i n i m u m of 12 independently controlled thrusters are needed (four for each axis, since two separate thrusters are needed for oppositely directed thrusts). For redundancy and greater simplicity of the controls, usually more are used. In m o s t cases these motors need to operate only outside the earth's atmosphere. Their nozzle expansion ratios can therefore be high, without
4.18 Liquid-Propellant Motors and Thrusters on Spacecraft
167
Figure 4.30 Bipropellant reaction control motor with ablation cooling. problems from overexpansion (Sect. 4.4.4). The larger motors are often capable of restart and throttling, whereas the reaction control thrusters are designed for a large n u m b e r of low-thrust, short pulses. Because of the dead period between pulses, the cooling requirements are less stringent than would be the case for continuous operation. Sufficient cooling, without the need for active cooling by convection, is obtained from radiation and from the ablation of the nozzle materials, such as carbon-phenolic or carboncarbon. In some cases, nozzles have also been manufactured from highly refractory metals, either tungsten or columbium. The pulse duration of reaction control thrusters may vary from a fraction of a second to as m u c h as minutes (although more rarely). Typical thrust levels are between 5 and 100 N. The total n u m b e r of pulses required during the life of a three-axis stabilized spacecraft may be as m u c h as 1 million or more. Because m a n y spacecraft, particularly communication satellites, require attitude and station keeping control during their entire useful life, which may be 10 years, the initial propellant mass becomes a major fraction ofthe spacecraft mass in some cases as m u c h as one-halfofthe total. Toreduce the requirement for propellant, bipropellants with their higher specific
168
CH A P T E R 4 Chemical Rocket Propulsion impulse are often preferred over the simpler monopropellants. However, when operating in short pulses, the specific impulse is significantly reduced compared with steady-state operation. The loss of enthalpy of the gas to the thrust chamber walls is no longer negligible, not only because of the small volume-to-surface ratio ofthese motors but also because at the beginning of each pulse the walls are still cold, hence the heat transfer high. Whereas the specific impulse of small hydrazine motors in continuous operation ranges from about 220 to 235 s, it may be only about 150 s when operated in short pulses. Because spacecraft rocket motors must reliably function even after many years of exposure to the space environment, thermal protection of the tanks, valves, and propellant lines is required. In m a n y cases electric heating must be provided to prevent freezing in deep-space missions or during solar eclipses. For the propellant in the weightless environment to cover the tank exit ports, the tanks must be equipped with bladders or diaphragms. Because of the small flow rates that are typical of reaction control systems, the need to cover the exit port can also be met by capillary feeds. An example is shown in Fig. 4.31, showing one of the propellant tanks of the NASA Space Shuttle reaction controls. Figure 4.32 illustrates a bipropellant reaction control system that uses high-pressure helium for the propellant feed. Each of the 16 motors is provided with a pair of injection valves, one for the fuel and one for the oxidizer. Together, the motors control the pitch, yaw, and roll of three-axis stabilized spacecraft. Redundancy is obtained by providing two parallel systems. Each can operate independently, yet, by command, the two systems can also be
Figure 4.31 Space Shuttle reaction control propellant tankwith capillary feed, NASA. (Courtesy of Lockheed Martin, U.S.A.) From Ref. 12, Brown, C. D., "Spacecraft Propulsion," AIAAEducation Series. Copyright 9 1996, AIAA~ reprinted with permission.
4.18 Liquid-Propellant Motors and Thrusters on Spacecraft
169
Figure 4.32 Schematic of a bipropellant reaction control system for a three-axis stabilized spacecraft. (Omitted are the fill and purge ports.) cross-linked so that the propellant of one system can also be used by the motors of the other. This type of redundancy is analogous to what is c o m m o n practice in the design of electronic systems. An important consideration is that crosslinking must not introduce n e w failure modes. In the schematic shown in the figure, this is accomplished by introducing multiple isolation valves. They make it possible, if needed, to separate the two sides of the system and also to isolate from each other the pitch, yaw, and roll motors. The valves are solenoid controlled and are either fully open or fully closed. They are governed by the spacecraft computer in response to the need for thrust or to redirect the propellant flow in case the pressure sensors indicate a failure in a part of the system. Such commands by the spacecraft computer can be overridden by ground command. Other features include pyrotechnically operated valves. They can be operated only once but have the advantage of being absolutely tight against leaks. They start the flow of the pressurization gas by bursting their diaphragm. This type ofvalve will stay leak free for many years, a feature that is
170
C H A PT E R 4
Chemical RocketPropulsion
particularly important when the system will not be needed until much later in a long-duration mission, as is the case with planetary descent or ascent vehicles. To guard against a pyrotechnic valve not opening on command, two parallel valves are provided on each side of the system. Because the helium pressure decreases as the propellants are being forced from their tanks, pressure regulators, again duplicated for redundancy, are located between the helium and propellant tanks. They reduce the pressure in the propellant tanks to the prescribed motor injection pressure. Backflow of helium or of propellant is prevented by check valves. These are arranged as quadruplets in an arrangement that is both in parallel and in series. (If one of the valves is stuck closed, the parallel string will still permit flow in the proper direction; if one does not close in spite of backflow, the valve that is in series will do so.) Burst diaphragms are placed on the helium side. They serve to prevent more serious damage if a pressure regulator should fail or if errors are committed in ground operations.
4.19
Components of Solid-Propellant Rocket Motors In solid-propellant motors the fuel and oxidizer, together with a binder and other additives, are premixed and contained as a solid mass in the motor case. The specific impulse is lower than that of most liquid-propellant motors, but this disadvantage is compensated by greater simplicity and lower cost. Solid-propellant motors are particularly well suited for boosters slung to the sides oflaunch vehicles and for rocket upper stages. When used as boosters, their lower specific impulse is acceptable because by their added thrust they reduce the gravity loss otherwise typical of massive launch vehicles.
4.19.1
Propellants So-called double-base propellants are a homogeneous mixture, typically a nitrocellulose explosive dissolved in nitroglycerin plus some small amounts of additives. Each of these components is both a fuel and an oxidizer (therefore the designation "double base"). They are explosive both separately and together. Composite propellants are mixtures of oxidizer crystals, often ammonium perchlorate, and a fuel in powder form, mixed with a binder such as polybutadiene. The processing is somewhat less hazardous than it is with double-base propellants. Also in use are combinations of these two types. They are then called composite double base. They consist of a crystalline oxidizer and a powdered fuel in a matrix of nitrocellulose-nitroglycerin. Among the additives are binders (which often also participate as fuels in the combustion), curing agents, and catalysts. The latter are important for improving the rheological properties of the mixture and for controlling the burn rate.
4.19 Components of Solid-Propellant Rocket Motors
171
At slightly elevated temperatures the premixed propellants turn into highly viscous fluids that can be cast into the motor case. After curing, and back at normal temperature, the propellants form a solid, rubber-like mass. Long-term storage requires a temperature-controlled, low-humidity environment. At low temperature, as may be encountered after an extended period in space, solid propellants may become brittle. Cracks or separation from the motor case may occur, probably with catastrophic consequences at the time of firing. The propellants therefore must be protected against the large temperature swings and the low temperatures frequently encountered after launch. The hazards encountered with solid propellants are fires and explosions, particularly in the processing and casting operations. These must be carried out with stringent controls. Once cast and cooled to normal temperature, the principal hazard is from electric sparks. As a precaution, the motor case and tools in contact with the propellant need to be electrically grounded. Oxidizers
The most frequently used oxidizer is a m m o n i u m perchlorate in the form of small, white crystals. It is chemically compatible with most fuels. Combined with high-energy fuels, it results in a high specific impulse. Like all other perchlorates, it produces hydrogen chloride in the combustion gas, rendering it toxic and corrosive. A m m o n i u m nitrate and other nitrates have the advantage of producing combustion gases that are smokeless and relatively nontoxic but result in a lower specific impulse in comparison with a m m o n i u m perchlorate. A m m o n i u m nitrate is often used in the gas generators that drive the turbo p u m p s of large liquid-propellant rocket motors. A high-performance oxidizer is s o d i u m perchloride. Its use as an oxidizer for rocket propulsion is limited by its hygroscopic nature. Fuels and Binders
Powdered a l u m i n u m has become an important fuel. Typically it constitutes up to 20% of the weight ofthe propellant. Upon combustion, it forms a slag of aluminum oxide, which is liquid at the motor temperature. Some fraction of it is retained in the motor case, particularly when reentrant nozzles are used. Retained slag does not contribute to the thrust, hence results in a (small) decrease of the specific impulse. Beryllium is much lighter than a l u m i n u m and, by comparison, can increase the specific impulse by about 15 s. Its use has been impeded by concerns about h u m a n health, because inhalation of beryllium powder is highly toxic. Beryllium hydride could increase the specific impulse still further by about 25 s. However, it deteriorates quickly through chemical reaction during storage and is therefore rarely used. Binders are most often polymers, such as several types ofpolybutadiene or plastisol binders such as polyvinyl chloride. They are essential to the safety and reliability of the motor. In the process of curing they determine the
172
C H A P T E R 4 Chemical Rocket Propulsion rheological properties of the final product by their cross-linking and branch chaining. The binders participate as a fuel in the combustion process and thereby add'to the enthalpy of the combustion gas. .
4.19.2
Burn Rate and Ignition The b u r n proceeds along the entire exposed surface, essentially at a spatially uniform rate. Given the initial propellant ("grain") surface, the b u r n surface at a later time into the burn can be found by applying at each time step a uniform recession distance perpendicular to the preceding surface. The b u r n rate, b, which is the rate at which the burn surface recedes, depends primarily on the gas pressure, Pch, in the c h a m b e r ("bore"). Because the Mach n u m b e r of the flow is low, in most cases Pch is essentially the same as the stagnation pressure. In turn, Pch is dictated by the instantaneous area of the burn surface relative to the area of the nozzle throat. The burn rate increases with pressure. Empirically it is found that it can be expressed by an exponential law of the form b - c (Pch/Pch) n
(4.56)
where n is referred to as the b u r n rate pressure exponent. Depending on the propellant, n can vary from as low as 0.2 to as high as 0.8. Propellants with n close to 1 are hazardous, because the steep increase of the b u r n rate with pressure can lead to a runaway condition of gas generation and pressure. It is convenient to choose for Pch in (4.56) a constant pressure, such as 100 atm, which is in the midrange of rocket motor pressures. (Different choices merely affect the factor c, not the burn rate pressure exponent.) The factor c depends strongly on the propellant. Adding a catalyst or reducing the oxidizer particle size can greatly increase c and therefore the b u r n rate. In addition, c depends somewhat on the initial propellant temperature. It is lower when the propellant is cold, say -20~ higher when warm, say 40~ This d e p e n d e n c e on temperature needs to be considered particularly w h e n the motor has been exposed for some time to the space environment, resulting in a cold propellant. As an example, for a composite a m m o n i u m nitrate propellant the burn rate at P c h ~ 50 atm is 1.7 m m / s at - 2 0 ~ and 2.1 m m / s at 40~ Figure 4.33, adapted from Sutton and Ross [1], provides additional data. It follows from the definition of the burn rate that the gas mass flow rate, r~, is given by /~1 - - A b 0 p r
b
(4.57)
where Ab is the burn surface area at some specified time and 0pr the propellant density. It is important to note that if there are cracks in the propellant or if the propellant mass has partially separated ("debonded") from the case wall, the effective b u r n area can be greatly increased over the intended area. Because the nozzle throat area is fixed, the increase in mass flow rate will increase the pressure in the motor case, possibly leading to a catastrophic failure. Cracks and separations are more likely to occur when the propellant in a spacecraft motor is very cold. T6 prevent this; the propellant temperature
4.19 Components of Solid-Propellant Rocket Motors I
I
I
I
I
I~
40
!
30
~!;4~
JPNu~ype
20
1
15
E E r~
173
-
66~ _
10
6.0
_
Composite Ammonium Perchlorate
m 4.0
60~ 3.0
-
2.0
m
1.5
Nitrate
1.0 I
20
I
I
I
I
I
I
30 40 60 100 150 200 Chamber Pressure, atm.
Figure 4.33 Burn rate versus chamber pressure for several solid propellants and initial propellant temperatures. (Adapted from Sutton and Ross [1],) needs to be controlled, either by thermal blankets or combined with electrical heating. Cracks and debonding can also occur after long-term storage in a less than perfectly controlled environment. Ultrasonic and x-ray probes are useful for prelaunch checks. Even in the absence of cracks there will be some unavoidable combustion instabilities. The use of powdered aluminum as a fuel additive has largely eliminated severe transverse oscillations. Slots or perforations in the propellant grain also tend to d a m p e n the pressure oscillations. It is often desirable to design the propellant configuration such that the pressure in the chamber stays approximately constant during the entire burn. The advantage is that the pressure can stay close to the m a x i m u m that is tolerated by the motor case. The result will be that the thrust is also nearly constant and a maximum for the case and nozzle design. In comparison to a thrust level that changes with time, yet has the same peak value, the gravity loss of the vehicle will be reduced. If the bore surface were simply cylindrical and the pressure constant, the burn area would increase linearly with time. Because the mass flow rate would increase, the pressure would in fact increase, resulting in a still more rapid increase of burn area, pressure, and thrust. This effect can be alleviated by designing the initial propellant inner surface to have, for instance, a
174
C HA PZ E R 4
Chemical Rocket Propulsion
star-shaped cross section, thereby increasing the initial burn surface without increasing the final one near burnout. For the same purpose, perforations or radial slots in the propellant can be provided (at the cost of a more complicated mandrel). To meet varying mission requirements, solid-propellant motors can be offloaded. This means that the motor is loaded with less propellant than could be accommodated by the case. Hence the motor is lighter and provides a smaller total impulse, as may be required for some missions. Ignition is initiated by squibs, which are small cylinders containing an explosive charge fired by an electric filament ("bridge wire"). For redundancy, some squibs have more than one filament. The squid starts an explosive transfer train that is similar to but m u c h faster than the type of powder cord used, for instance, in mining. In turn, the transfer assembly ignites the propellant. Every possible precaution must be taken to prevent premature ignition. A typical arrangement is that in the safe position the explosive transfer train is physically separated from the squib. Arming then consists of eliminating this separation by the mechanical action provided by a small electric motor. All electrical parts that form part of the ignition system are designed such that at least two independent failures would have to occur before a premature ignition becomes possible.
Figure 4.34 Solid-propellant boosters on the U.S. Space Shuttle, early design. (a) Stack of the segments, indicating locations of field joints; (b) design of field joints prior to Challenger accident.
4.20
4.19.3
Hybrid-Propellant Rocket Motors
175
Case Design Usually, the motor case is of approximately cylindrical or spherical shape. Smaller, cylindrical cases can be fabricated as a single piece ("monolithic design"). Steel or composites are commonly used. For very large motors, such as the solid-propellant boosters attached to major launch vehicles, the difficulty of casting the propellant and transporting the motor to the launch site requires a buildup of a stack consisting of separate segments. The case walls of the segments are connected by bolts or by tang and clevis held in position by pins. An example of the latter design is the boosters on the U.S. Space Shuttle. Figure 4.34 shows an earlier version, now superseded. The figure indicates some ofthe precautions n e e d e d to mate the individual segments properly. Insulation and flame-inhibiting materials are used to prevent the flame from spreading into the interface between adjoining segments. To ensure good bonding to the case, a layer of inert, organic material is used. The same material also acts as a thermal insulator to the case w h e n the propellant is close to being c o n s u m e d at any one location. The joints between segments are designed to be leak tight by the use of rubber O-rings in series, protected from the hot gas by a zinc chromate putty.
4.20
Hybrid-Propellant Rocket Motors In these motors the fuel is solid, in a configuration that is similar to that of solid-propellant motors. In contrast, the oxidizer is liquid, stored in a tank and p u m p e d to impinge on the b u r n surface of the fuel. In terms of simplicity of the design and usually also in terms of the specific impulse, hybrids are intermediate between solid- and liquid-propellant motors. They have the ability to be throttled, to be extinguished, and to be restarted. An example is provided by a class of motors that use hydrogenterminated-polybutadiene (HTPB) as the fuel and N20 as the oxidizer. The turbine that drives the oxidizer p u m p is provided with gas from a gas generator that makes use of the same oxidizer and of an azide-based polymer solid fuel.
Nomenclature a
b Cp, Cv e
g h m m
speed of sound burn rate (Eq. 4.56) specific heats internal energy gravitational acceleration; go: standard gravitational acceleration enthalpy mass mass flow rate
176
C H A P T E R 4 Chemical Rocket Propulsion ni P r
rm S U~ Y~ W X
xi Y A
Ft Isp K(T)
M, Nu, Pr, Re R
T or
flex E
/z /xj vi
v(M) 0 o)
()a
()e ( )ch ( )ex
(). ( )pl ( )pr
( )rs ()vac ()*
n u m b e r of moles of the ith constituent pressure radius mixing ratio entropy velocities coordinate mole fraction of ith constituent control variable area thrust specific impulse equilibrium constant (Eq. 4.41) Mach, Nusselt, Prandfl, Reynolds n u m b e r s gas constant absolute temperature constant in Summerfield criterion (Eq. 4.25) divergence angle staging residual mass fraction; also degree of reaction Lagrange multiplier Mach angle; also viscosity molecular weight ofjth constituent stoichiometric coefficient of ith constituent Prandtl-Meyer function density sequential mass fraction; ~: payload mass fraction angular velocity ambient azimuthal chamber nozzle exit plane meridional payload propellant residual v a c u u m condition condition at throat
Problems (1) Consider the nozzle represented in Fig. 4.35. An axisymmetric bell nozzle with zero divergence angle and the dimensions (in mm) indicated in the figure is assumed. There is zero ambient pressure at the nozzle exit. The calculation can be based on the approximation of a steady, onedimensional, isentropic flow of a thermally and calorically perfect gas with an average ratio of 1.30 of the specific heats and a gas constant of 330 m2/(s 2 K). The stagnation temperature can be taken equal to the flame
Problems
177
/ /
I I
2400 R
200 D -
480 ~
1520
Figure 4.35 Axisymmetric bell nozzle. (Dimensions are in mm.)
temperature of 3200 K in the combustion chamber. The stagnation pressure, which can be taken equal to the c h a m b e r pressure, is 800 N ! cm 2. (a) Find the Mach n u m b e r at the nozzle exit. (b) Find the pressure, density, temperature, and velocity at the nozzle throat. (c) Find the mass flow rate, the theoretical thrust (including the "pressure thrust"), and the specific impulse. (d) Graph as functions of the axial coordinate the Mach number, the temperature, and the pressure. (2) Consider a solid-propellant motor with the grain cross section indicated in Fig. 4.36. This cross section can be assumed to be uniform over the length of 1500 m m of the grain. The density of the propellant is 1700 k g / m 3. The nozzle throat area is 4000 m m 2, the nozzle exit area 0.300 m 2. The motor operates in vacuum. The flame temperature in the gas passage perforation is 3300 K and can be taken equal to the stagnation temperature of the flow. A perfect gas is assumed with ratio of the specific heats of 1.25 and a gas constant of 280 m 2 / (s2 K). (a) Given a constant b u r n rate of 3.5 m m / s of the propellant, show graphically the inner contour of the propellant at 0, 20, 40, 60, and 80 seconds after ignition and find the mass flow rate at these times. (b) Find the stagnation pressure (= pressure in the center perforation) at these times. (c) Find the thrust (including the "pressure thrust") at these times. (d) From the b u r n rate of 3.5 m m / s at ignition and a b u r n rate pressure exponent of 0.6, find improved values for the b u r n rates at 20, 40, 60, and 80 seconds by using the results obtained in (b).With these improved numbers, recalculate the thrust at these times. (3) For the rocket m o t o r defined in Problem 1 and using the results obtained there, find (a) The ambient pressure at which there will be a normal shock at the nozzle exit plane.
178
C H A P T E R 4 Chemical Rocket Provulsion
Figure 4.36
Cross section of a solid-propellant grain. (Dimensions are in mm.)
(b) The ambient pressure above which flow separation in the nozzle is likely to occur. [Make use of the Summerfield criterion, Eq. (4.25).] (4)* Consider a propellant p u m p designed to p u m p liquid oxygen (density 1.19 103 k g / m 3 at a mass flow rate of 35.0 kg/s. The p u m p is ofthe centrifugal type, similar to the one shown in Fig. 4.20. The flow enters the p u m p in the axial direction, then flows radially outward, driven by the impeller blades. Finally, the flow is redirected by the stationary blades in the volute such as to leave the p u m p in a conduit without swirl. To simplify the calculation, an infinite n u m b e r of blades are assumed, so that the flow is steady and independent of the azimuthal coordinate. To further simplify, the entire flow is represented by the flow tangential to the m e a n flow surface (defined as the axisymmetric surface for which one-half of the mass flow is on the inside, one-half on the outside). Losses will be neglected. At the p u m p entrance, the hub radius is 30 mm, the blade tip radius 100 mm. The impeller exit radius is 150 mm. The widths of the flow passages are such that the throughput velocity (i.e., the velocity c o m p o n e n t in the meridional planes) is the same everywhere. The impeller blades are curved and turn the flow relative to the impeller through an angle of 60 ~, measured along the m e a n flow surface. The rate of rotation of the impeller is 7000 rpm. The oxygen upstream of the p u m p is at a temperature o f - 2 2 0 ~ and a pressure of 2.00 atm, which corresponds to a boiling temperature o f - 176~
References
179
(a) Compute the pressure at the impeller exit and the pressure at the p u m p exit. (b) Compute the theoretical (no loss) power required to drive the pump. (5)* A space vehicle that, in addition to the payload, has two rocket stages is assumed to accelerate in gravity-flee space along a rectilinear path. The specific impulse for the lower stage is 350 s, for the upper stage 440 s. For both stages, the stage residual mass fraction is 0.12. The ratio ofpayload mass to total takeoff mass of the vehicle is 0.10. Using the method of Lagrange multipliers, compute the o p t i m u m ratio ofthe first-stage mass (including its propellant) to the takeoffmass. Similarly, compute the ratio of the second-stage mass to the takeoff mass. (Answers: 0.522 and 0.377.) (6)* Consider the chemical reaction 3S_2+S'1
<
~
2S1 +$2
where the Si designate the participating species. The degree of reaction will be designated by e (referred to the product species $1), the ratio of the pressure to a standard pressure by p/Po, and the equilibrium constant by K(T). Write the mass action law in the form
f(e, P/Po) = logK(T) by expressing the function f explicitly in terms of e and (p/Po). (7)* Assume a rocket motor with a fission reactor and hydrogen as the propellant. The hydrogen cools the reactor and enters a nozzle for expansion to a supersonic velocity. The hydrogen will be partially dissociated. For an assumed temperature of the hydrogen at the nozzle entrance of 3000 K and a pressure of 30 atm, compute the degree of reaction pertaining to the dissociation to atomic hydrogen. (The equilibrium constant for this reaction, at 3000 K, is 2.838 10-2; the reference pressure is 1 atm.)
References 1.
Sutton, G. P. and Ross, D. M., "Rocket Propulsion Elements," 5th ed, John Wiley & Sons, New York, 1986.
2.
Huzel, D. K. and Huang, D. H., eds., "Design of Liquid-Propellant Rocket Engines," American Institute of Aeronautics and Astronautics, Washington, DC, 1992.
3.
Vertregt, M., 'Tk Method for Calculating the Mass Ratios of Step Rockets," Journal of the British Interplanetary Society, Vol. 15, pp. 95-97, 1956.
4.
Liepmann, H. W. and Roshko, A., "Elements of Gasdynamics," Galcit Aeronautical Series, John Wiley & Sons, New York, 1957.
5.
Eyring, H., Lin, S. H. and Lin S. M., "Basic Chemical Kinetics," John Wiley & Sons, New York, 1980.
6.
Hansen, C. E, "Rate Processes in the Gas Phase," NASA Reference Publication 1090, NASA, Washington, DC, 1983.
180
C H A P T E R 4 ChemicalRocketPropulsion 0
,
,
10. 11.
12.
Wilkins, R.L., "Theoretical Evaluation of Chemical Propellants," Prentice-Hall, Englewood Cliffs, NJ, 1963. Glassman, I. and Sawyer, R. E, "The Performance of Chemical Propellants," NATO Advisory Group for Aerospace Research and Development, AGARDograph No. 129, Technivision Services, Slough, England, 1969. Vance, J. M., "Rotordynamics of Turbomachinery," John Wiley & Sons, New York, 1988. Brennen, C. E., "Hydrodynamics of Pumps," Oxford University Press, Oxford, 1994. Oppenheim, B. W., and Rubin, S., 'Tkdvanced Pogo Stability Analysis for Liquid Rockets," Journal of Spacecraft and Rockets, Vol. 30, No. 3, pp. 360-373, 1993. Brown, C. D., "Spacecraft Propulsion," AIAA Education Series, American Institute of Aeronautics and Astronautics, Washington, DC, 1996.
Orbital Maneuvers In this chapter, we will consider some special flight paths that are of particular interest for a n u m b e r of space missions, as described in Refs. 1 to 4. As before, attention will be restricted to flight paths, or segments of such paths, where the gravitational field is a central, inverse-square force field. Perturbations, such as those that are produced bythe oblateness ofthe earth, will therefore be neglected. In some cases, the spacecraft center of mass motion that is considered here will be one of merely coasting (i.e., in free fall) without thrust. Other cases that are considered are flight paths where intervals of coasting are separated by short, quasi-instantaneous bursts of thrust. Here, the thrust vector, pointing in a fixed direction, is represented as a delta function of time. In m a n y cases, this turns out to be a good approximation because the thrusting time (the so-called burn) of chemical rocket motors is usually just a few minutes or less, m u c h shorter than orbit periods. Whereas most examples in the present chapter will apply to artificial satellites and other spacecraft, in a later section, "gravity turns," maneuvers typical for m a n y launch vehicles, will be considered.
5.1
Minimum Energy Paths In the planning of space vehicle trajectories, it is often of interest to find the m i n i m u m energy p a t h that leads from a prescribed point of departure to a prescribed target point. Both points, as well as the center of attraction, are assumed to be fixed in the same inertial reference system, for instance, the heliocentric system. Because by (3.12) the energy is directly related to the semimajor axis, the problem becomes a purely geometrical one. It is illustrated in Fig. 5.1 for the case of elliptic orbits. Given a point of departure P1 and target point P2, there are infinitely m a n y arcs of conic sections that connect the two points and are in the plane containing the center, designated by F, of the gravitational attraction. For a fixed trajectory plane and a fixed center of attraction in it, the n u m b e r of parameters needed to determine the trajectory would be three: the semimajor axis a, the eccentricity e, and an angle to fix the inclination of the major axis. If, in addition, two distinct points, such as P1 and P2, through which the trajectory must pass are specified, the n u m b e r of free parameters, in the general case, is reduced to one. In the figure, two of the infinitely m a n y possible ellipses through P~ and P2 with center of attraction F are shown. The two ellipses have different 181
182
C HA PT E R 5
Orbital Maneuvers
Minimum Energy Path
Figure 5.1 Elliptic trajectories connecting two fixed points P1 and
Pz
in an inverse-square gravitational field.
secondary foci, designated by F' and F m, respectively. The triangle P1-Pa-F is given by the data, but the secondary foci are as yet unknown. For each ellipse, there are two c o m p l e m e n t a r y arcs that connect the point of departure P1 to the target point P2. The energy, designated by w (potential plus kinetic energy) per unit mass of the space vehicle, is a constant of motion, the same for both arcs. However, in general, the flight times will differ. From a well-known geometrical property of ellipses, the radii indicated in the figure must satisfy the relations !
l
rl + r 1 = r2 + r e = 2 a
(5.1)
It follows from r'a = 2a - ra and r~ - 2a - r2 that such elliptic arcs always exist irrespective of the choice of P1, P2, and F, provided only that the semimajor axis is chosen large enough. For then the two circles of radius r'1 about P1 and radius r~ about P2 will intersect. Let d designate the distance P~-P2 and s = (1/2)(rl + r2 + d), hence one-half of the perimeter of the triangle P1-Pz-E From the triangular inequality r'1 + r; >__d a n d (5.1) follows 2a >__s. Therefore, for a prescribed triple ofpoints P1, P2, and F, the m i n i m u m value, am, that the semimajor axis can have is 1 am -- gS
The radii r!lm and
r~m
(5.2)
associated with this m i n i m a l ellipse b e c o m e !
rim
!
-- s -
rl,
r2m =
S - - /'2
(5.3)
and therefore !
!
rim + r2m
--
d
(5.4)
This shows that the secondary focus F m of the minimal ellipse lies on the line PI-P2, with the distance d divided according to (5,4).
183
5.1 Minimum Energy Paths For the minimal ellipse, the energy/13 m per unit mass from (3.12) is
(5.5)
Wm = --IZ/S
Since the energy decreases with decreasing semimajor axis, the minimal ellipse is also the path from P1 to P2 that requires the least energy. It also follows that the minimal ellipse corresponds to the smallest kinetic energy at the point of departure, hence to the m i n i m u m initial velocity. One concludes from (3.14) that this velocity (in a nonrotating reference frame attached to the center of attraction) is
Vlm-- ~2/Z(~l -- l )
(5.6)
It should be recalled, however, that the preceding development applies only to cases where the points P1 and P2 are fixed in the nonrotating reference flame attached to the center of attraction. This is the case, for example, for the exoatmospheric path of a rocket or for the aircraft-launched suborbital flight of a vehicle, but not for a mission from the earth to a planet, because in this last case neither the point of departure nor the target is fixed. Nevertheless, even in this case, the preceding equations are often useful for estimating purposes. The eccentricity, era, of the minimal ellipse can be found as follows. Applying to the triangle P1-Pa-F the identity that relates for any triangle one of its half-angles to its perimeter and the lengths of its sides, one finds
az
/s(s
COST
V
rl) r2d
where a2 is the angle formed by 1"2and r~m. The distance between the foci F and F m is 2 am em = Sem. Therefore from the trian gl e F m-P 2-F see m =
a +
- 2
= (s-r2) 2+r 2 _s2
4s(s
mr2 c o s t a
2(s-r2)r2
rl)(S d
2S(S rl) _ 1 r2d
E
J
1"2)
hence
em
~
/1 - 4(s-
r~)(s- r2) sd
(5.7)
Having found the semimajor axis and eccentricity of the m i n i m u m energy path, the flight time between the point of departure and target can be calculated from Kepler's equation (3.21). Although straightforward, the calculation is algebraically c u m b e r s o m e unless Lambert's theorem, discussed in the next section, is used. Results analogous to those derived for elliptic paths can also be found for hyperbolic paths. The energy, w, in this case is always positive, hence
184
C HA PT E R 5
OrbitalManeuvers
larger than the energy required for an elliptic path through the same pair of points.
5.2
Lambert's Theorem Lambert (1728-1777) established the surprising fact that in a central inversesquare force field the flight time between two points (fixed, as before, in the nonrotating reference frame attached to the center of attraction) is proportional to the Kepler orbit period, with a proportionality constant that depends only on the straight line distance between the points and the sum of the distances of the points from the center of attraction. Figure 5.2 illustrates the theorem and indicates the notation that will be used. The proof of Lambert's theorem makes frequent application of the identities that relate sums of trigonometric functions to their products and vice versa. The theorem is valid for all conic sections but will be derived here only for elliptic paths. The time of flight from point P1 to point P2 as given by Kepler's equation (3.21) is t2 -- tl = =
(P/2yr)[E2 (P/27r)[E2-
-
E1 -
e(sin E2 - sin El)]
1 E1 - 2 e c o s ~(E1 + E2)sin ~1 ( E 2 - El)]
(5.8)
where P is the period of the orbit. This form is useful because it shows that the flight time depends, other than on the period, only on the sum and difference of the two eccentric anomalies E~ and E2. In place of E1 and E2 one specifies new variables f and g by the two equations f + g - 2 cos - 1 (e cos ~1 (El + f - g =
E2))
E2-E1
Figure 5.2 Illustration of Lambert's theorem.
(5.9)
5.2
Lambert's
185
Theorem
(We p o s t p o n e until later the c o n s e q u e n c e s of the m u l t i v a l u e d n e s s of the arc cos function.) F r o m s u b s t i t u t i o n into (5.8), t2 -- tl - -
- sin f -
(P/2~)[f
( g - sing)]
(5.10)
The same functions f a n d g can also be u s e d to express the distance d b e t w e e n P~ a n d P2. For, if (Xl, Yl) a n d (x2, Y2) are the Cartesian coordinates, as s h o w n in the figure, o f P1 a n d P2, respectively, d2 -
(x2 - x1) 2 -~- (Y2 - Yl) 2 - a 2 ( c Os E2 - c o s E l ) 2
+ a2(1 - e2) (sin E2 - sin El) 2 1 1 1 = 4a 2 sin 2 ~(E2 - El)[ sin 2 ~(E1 + E2) + (1 - e 2) cos 2 ~(E1 + E2)] _ g). s i n 2 l ( f
= 4 a 2 s i n 2 ~l ( f
+ g)
from (3.19) a n d the s u b s t i t u t i o n s from (5.9). F r o m the m a t c h i n g trigonometric identity it follows therefore that d - a(cos g - cos f )
(5.11)
An equally simple expression can also be o b t a i n e d for the s u m rl + r2 of the two radii. Using the e q u a t i o n following (3.19) a n d applying twice the trigonometric identities relating s u m s a n d products, rl+r2
=2a[1-
1 e(COS E1 + COS E2)]
= 2aIl - c~
f + ] g c2~
f 2- g
= a[2 - cos f - cos g]
(5.12)
Finally, solving (5.11) a n d (5.12) for cos f a n d cos g, one obtains cos f - 1 - (2a) -1(rl + r2 + d) cos g - 1 - (2a) -1 (rl + 1"2 - d)
(5.13)
The ambiguity, n o t e d earlier, in solving for f a n d g can be resolved as follows: We define f * a n d g* as the principal values of f a n d g, so that fk, l . . . . .
+ f * + 2zrk,
g-
+g* + 2zrl
(5.14)
1, O, 1, 2, . . . . Since d >_ O, it t h e n follows from (5.13) that f * _> g*
(5.15)
Also, (5.10) can be rewritten as t2 - tl - ( P / 2 n ) [ + ( f *
- sin f*) T (g* - sing*) + 2zrm]
(5.16)
where m - 0 or 1 a n d the choices o f t h e algebraic signs o f t h e first a n d s e c o n d s u m m a n d s are as yet i n d e p e n d e n t of each other. (We exclude orbits consisting of m o r e t h a n one revolution; their flight times are o b t a i n e d trivially by a d d i n g integer multiples of the Kepler orbit period.) Four cases, r e p r e s e n t e d in Fig. 5.3, n e e d to be distinguished. They dep e n d on w h e t h e r the area A (to the left in m o v i n g from/)1 to P2) b o u n d e d by the a r c P 1 - P 2 a n d the line P 1 - P 2 does or does n o t c o n t a i n the focal points F and F'.
186
C H A PT E R 5
Orbital Maneuvers
Figure 5.3
The four cases of Lambert's theorem.
Case (a): H e r e n e i t h e r F n o r F' is in A. T h e v a l u e of m a n d t h e algebraic signs in (5.16) c a n m o s t easily be f o u n d f r o m t h e l i m i t i n g case w h e n P1 a n d P2 coincide, t h a t is, w h e n d = 0, t2 - tl = 0. Clearly, ra = r2 < 2a. H e n c e f r o m (5.13) cos f = cos g 7~ - 1 a n d f * < zr, g* < n. By continuity, also in t h e g e n e r a l case, m = 0 a n d
t2-ta = ( P / 2 n ) [ + ( f * - s i n f * ) - ( g * - s i n
g*)]
i f F r A, F' r A
(5.17a)
since, by a s s u m p t i o n , t2 - t~ >_ 0 a n d f * - sin f * >_ g* - sin g*. T h e limit of case (a) is r e a c h e d w h e n F' is o n t h e line P~-P2. In t h a t case, t d = r~! + r2. F r o m rl + r]t - 1"2 + r~ - 2 a a n d (5.13) follows t h a t cos f - - 1 , h e n c e f * - sin f * = Jr at t h e t r a n s i t i o n f r o m case (a) to case (b). Case (b) o c c u r s w h e n F', b u t n o t F, is in A, e a s e (c) w h e n b o t h F a n d F' are in A, a n d e a s e (d) w h e n F, b u t n o t F', is in A. F i n d i n g m a n d t h e algebraic signs in (5.16) in e a c h of t h e s e cases p r o c e e d s a n a l o g o u s l y to case (a). T h e results c a n be s u m m a r i z e d as follows: t2 - tl = (P/2n)[2zr - ( f * - sin f * ) - (g* - sing*)]
ifFeA,
F' e A
(5.17b) t2 - tl = (P/2zr)[2n - ( f * - sin f * ) + (g* - sing*)]
ifF~A,F'
eA
(5.17c) t2 - tl = (P/27r)[+( f * - sin f * ) + (g* - sing*)]
ifF~A,F'
CA
(5.17d) L a m b e r t ' s t h e o r e m , w h i c h h a s j u s t b e e n derived, is similar to Kepler's e q u a t i o n i n s o f a r as it allows o n e to c o m p u t e f r o m orbital d a t a t h e e l a p s e d t i m e of flight b e t w e e n two points. T h e i n v e r s e p r o b l e m , the c a l c u l a t i o n of t h e l o c a t i o n of a s p a c e o b j e c t at a given time, u s u a l l y r e q u i r e s a n u m e r i c a l i n v e r s i o n or else a series e x p a n s i o n as d i s c u s s e d in Sects. 3.3.1 a n d 3.3.2.
5.2 Lambert's Theorem
5.2.1
187
Application to Suborbital Flights Lambert's t h e o r e m provides a particularly convenient way to calculate the exoatmospheric flight time of suborbital vehicles b e t w e e n two points. We neglect here for simplicity the corrections introduced by a e r o d y n a m i c drag, earth rotation, and gravity anomalies. Assuming that the chosen p a t h is a m i n i m u m energy trajectory (Section 5.1), the secondary focus, F', coincides with the line connecting the two points. The limiting case of the transition from case (a) to case (b) therefore applies, with f * - sin f* = zr. (The comp l e m e n t a r y path would lie in the earth's interior.) From (5.13), defining the constant S by =
2(rl + r2 - d)
(5.18)
rl + r2 + d
follows cos f = - 1 ,
cosg = 1 - 6
Therefore the elapsed flight time for the m i n i m u m energy p a t h becomes, from (5.17a) or from (5.17.b), t2
t~ -- (P/2Jr)[Tr - cos -~ (1 - 6) + ~/1
(1 - 6) 2]
(5.19)
where the arc cos a n d the square root functions are evaluated on their principal branches. The period, P, of the complete orbit, from (5.2) is ~ (rl + 1"2+ d) 3
5.2.2
(5.20)
Application to Planetary Missions Spacecraft paths from the earth to the planets or to the moon, or b e t w e e n planetary flybys, are calculated by highly accurate numerical methods. These are outside the scope ofthis book. The m e t h o d s outlined in Sects. 5.1 a n d 5.2, if iterated, are useful, however, for purposes of preliminary flight planning. We assume here that the spacecraft is already outside the sphere of influence of the earth and has a velocity sufficient to reach the planet on a m i n i m u m energy path without further use of thrust. In the heliocentric reference flame, an iterative s c h e m e can be developed as follows: Starting with the earth's position (point Pa) at the time of departure of the spacecraft and a rough estimate of the planet's position (point P2) at the time of arrival of the spacecraft, an estimate for the radii rl, r2 a n d the distance d b e t w e e n P1 and P2 is obtained. Lambert's theorem then allows one to calculate a first estimate of the flight time for a m i n i m u m energy path from Pa to P2. From this, a second, improved, estimate of the planet's position at the t i m e of arrival can be obtained. In turn, from Lambert's theorem, a second approximation is obtained for the flight time. Still better results will generally be obtained by continuing the iteration.
C H A PT E R 5
188
5.3
OrbitalManeuvers
Maneuvers with Impulsive Thrust Injection of a spacecraft into a final orbit or hyperbolic trajectory almost always consists of one or more lengthy coasting intervals separated by m u c h shorter intervals of thrust. Chemical propulsion lends itself to short thrust durations ("burns"). They can be of the order of minutes or less. For small, tactical missiles with high acceleration, the thrust duration may be of the order of only seconds. For the spacecraft maneuvers considered in this section, good approximations can be obtained by assuming that the thrust has a uniform direction and is impulsive, that is, represented by a delta function in time. The result is a change in speed
Av = ~i t~ F(t) dt t, m(t)
(5.21)
m
where Ft (t) is the thrust, re(t) the vehicle mass, and the intended limit is obtained by letting t a - t~ ~ O, Ft ~ oo. The m a n e u v e r is carried out by first orienting the vehicle so that its thrust axis is in the direction of the desired velocity change. Sometimes the objective of the maneuver is to produce a change in the orbital plane. At other times, the objective may be to increase the magnitude of the velocity. Often, these are combined by the application of a single, impulsively applied thrust.
5.3.1
Plane Changes Figure 5.4 illustrates a simple plane change from one circular orbit to another such orbit. The two orbital planes are defined by their inclinations ia a n d / 2 to the equatorial plane and the right ascensions f21 and ~2 of the ascending nodes. The speed, v, in this simple case is assumed to be
Figure 5.4 Plane change by impulsive thrust from a circular orbit to another circular orbit.
189
5.3 Maneuvers with Impulsive Thrust
unchanged. With V1 designating the velocity immediately preceding and v2 the velocity immediately following the thrust, there is vl -- v2 = v. The turning angle fl is obtained from cos fl - cos il cos i2 + sin il sin/2 cos(f22 - f21)
(pure plane change)
(5.22) as follows easily from the trigonometric relation relating the angles il, 7r - i2, a n d ~ 2 - ~1 in the spherical triangle shown in the figure. In this case, the two velocity vectors form an isosceles triangle. Hence the magnitude A v of the Av vector that must be p r o d u c e d by the thrust is Av -- 2v sin (2!fl)
(pure plane change)
(5.23)
Plane changes are needed, for instance, w h e n injecting a geostationary satellite or other equatorial satellite from a launch site not on the equator. As indicated in Fig. 5.5, the ground trace of the satellite follows at first a great circle through the launch site. A plane change can be m a d e at either point I or at point I' in the diagram, that is, on the line of intersection of the initial and final orbit planes. In the example, case (a), shown in the figure, a launch is assumed from the U.S. launch complex at Cape Canaveral (latitude 28.7 ~ north) with a launch direction approximately east (to take advantage of the earth's rotation and to ensure reentry of the launch boosters over the ocean). Considered here is the plane change from a nonequatorial, circular orbit into an equatorial circular orbit of the same radius and equatorial crossings at points I and I' located at longitudes 90 ~ and 270 ~ b e y o n d the longitude of the launch site. As is easily seen from the right angle spherical triangle shown in the figure, the turning angle/~ is just equal to the latitude ~L of the launch site. Therefore, from (5.23), A v / v = 2 sin(3L/2) = 0.496 for Cape Canaveral. For the European Space Agency, French-operated launch site (latitude approximately 5 ~) in Guiana, the corresponding n u m b e r is 0.087.
Figure 5.5 Ground traces for launch from Cape Canaveral, Florida (~ = 28.3 ~ for insertion into equatorial orbits: (a) from a circular orbit and (b) from a suborbital trajectory.
190
C H A PT E R 5 OrbitalManeuvers Another insertion A v can be obtained by first launching into a suborbital trajectory with a heading in a more southerly direction, case (b) in the figure. The transition into the final, circular, equatorial orbit in general will then be a c o m b i n a t i o n of plane change a n d speed increase. For a given turning angle fl, Av is less at lower speed v, indicating that pure turning is advantageously carried out for circular orbits at higher altitude and for elliptic orbits at the apoapsis.
5.3.2
Repeated Thrusts at Periapsis For a given initial velocity V1 and a required final velocity re, the needed velocity i n c r e m e n t A v, relative to Vl is AI)/U1
--
v/1 + (ve/va) 2 - 2 ( v z / v 1 ) cos fl
(5.24)
where fl is the angle between Vl a n d v2. If fl = 0, this ratio is a m i n i m u m . As an example, we consider the effect of repeated impulsive thrusts at the periapsis of elliptic orbits. Figure 5.6 illustrates the effect of separate firings, at the same point, of u p p e r stage motors, typically solid-propellant motors, of a multistage vehicle. The motor firings occur at periapsis, therefore where the velocity is highest for each elliptic path. Before each firing, the attitude of the spacecraft is readjusted to ensure that the thrust at the periapsis is in the
Figure 5.6 Repeated upper stage firings at a common periapsis.
191
5.3 Maneuvers with Impulsive Thrust
(a)
(b)
(c) V)2
,..V1
~'~
("~V)I~/// V1
J ~,~
Z~V)2 /
(AV)1
V1
Figure 5.7 Comparison of one versus two impulsive applications of thrust. direction of the path. Each elliptic path returns to the same periapsis height, with a velocity that is larger than for the preceding path. The elapsed time for each elliptic path allows enough time for stage separation and attitude adjustment for the next firing.
5.3.3
Plane Change and Speed Change Combined It is often advantageous to c o m b i n e plane changes with speed changes. Figure 5.7 represents the plane that contains the initial velocity v~ and final velocity v2. It shows three different modes for comparison. Often, the twin objectives of plane change and speed increase can be achieved with a single motor firing, case (a) of the figure. It is of interest to compare this with the two other cases: case (b) to make the plane change first, to be followed (still at essentially the same point on the trajectory, i.e., at the same potential energy) by increasing the speed, and case (c) to increase the speed first and then make the plane change. It follows from the triangular inequality applied to cases (b) and (c) that the required Av is smallest in case (a), that is, w h e n the plane change and speed increase are combined in a single firing.
5.3.4
Multiple Stages with Impulsively Applied Thrust In practice, these theoretically optimal combinations for achieving a mini m u m A v for a given objective can serve only as a general guide. This is because mission designs are necessarily almost always based on the use of already developed and available rocket motors. Whereas some liquidpropellant motors are capable of multiple starts and terminations, this is not the case for solid-propellant motors, the class of motors (possibly offloaded) that is most frequently used for maneuvers as described here. The mission designer therefore must typically choose from a relatively small set of existing, or at least planned, motors and from several possible choices a near-optimal solution. In what follows, we assume a two-stage vehicle. The motors are fired in quick succession, so that the gravitational potential energy is unchanged. The choice of rocket motors and stage characteristics then usually consists of maximizing the increase in the vehicle's total energy obtained by firing the two stages. (Extensions to more than two stages, and to a delayed firing of the second stage, hence at a different potential energy, are immediate.)
192
C H A PT E R 5 OrbitalManeuvers Let ml -- mass of stage 1 before b u r n ma = mass of stage 2 before b u r n m3 --- spacecraft mass #1 = stage 1 ratio of residual mass to mass m~ #2 -- stage 2 ratio of residual mass to mass m2 v0 = velocity before stage 1 b u r n (A W)3 = energy increase of the spacecraft, resulting from the stage firings The residual masses referred to comprise such m o t o r c o m p o n e n t s as the void m o t o r case, nozzle, igniter, and thrust cone of solid-propellant motors. Typically, the residual mass is no more than about 10% of the total stage mass before firing. The increase in the spacecraft's energy, (A W)3, is 1 ( A W ) 3 = ~m3(v0 q- ( A u ) I -~- (AI))2) 2
(5.25)
where (Av)~ and (Av)2 are the velocity increments provided by stages 1 and 2, respectively. From the "rocket equation" (2.28), they are ( A v ) I ~-
g0Isp,1 In
(Av)a = g0/sp, a In
m~ + m2 + m3 #a ma + ma + m3
(5.26a)
ma + m3 /zama + m3
(5.26b)
where Isp, a and Isp,a d e n o t e the specific impulse of the stage 1 and 2 motors. As far as propellant c o n s u m p t i o n is concerned, there is no difference w h e t h e r the residual of the last stage remains attached to the spacecraft or is separated from it. Separation m a y be required to prevent contamination of the spacecraft from outgassing that occurs from m o t o r s that are still hot. The mission designer's task is therefore to select a c o m b i n a t i o n ofrocket motors that will result, at least approximately, in the smallest c o m b i n e d mass of the two stages for the required spacecraft energy i n c r e m e n t A W3.
5.4
Hohmann Transfers A frequently executed m a n e u v e r is to raise a spacecraft from an initial, often circular, orbit into a higher, again often circular, orbit. The two orbit planes m a y or m a y nor coincide. This m a n e u v e r can be accomplished by two impulsive thrusts, one on the initial orbit, the other ("orbit insertion") immediately preceding the final orbit. Such a m a n e u v e r is called a H o h m a n n transfer.
In what follows, we consider the transfer of a spacecraft between two circular orbits, not necessarily in the same plane. One instance of this would be the transfer from a low earth orbit, with its orbital plane through the launch site, into a geostationary orbit. The m a n e u v e r is illustrated in Fig. 5.8. The initial orbit has radius ra and velocity 1)1. Similarly, the final orbit has radius r2 and velocity v2. The initial and final orbit planes a n d the transfer
5.4
Hohmann Transfers
193
Figure 5.8 Hohmann transfer between two circular orbits; P, periapsis, A, apoapsis. orbit plane are a s s u m e d to intersect all in a c o m m o n line, referred to as the line of absides. The first thrust changes the velocity vl to v~ t h r o u g h the angle fl~ so that the n e w velocity is in the H o h m a n n transfer plane. Similarly, the second thrust changes the velocity v~ to v2 t h r o u g h the angle f12 so that the new velocity is in the final orbit plane. The transfer orbit therefore is a semiellipse with its periapsis on the initial orbit and its apoapsis on the final orbit. It follows that the s e m i m a j o r axis, a', of the transfer orbit is a ' = (1/2) (r~ + r2). Therefore, from (3.14), 1)1 =
v2 = 4 ~ / r 2
x / lz / r l ,
,
v! _ / 2/~r2 1 V rl (rl + rz)'
~
1)2 =
(5.27)
21xrl r2 (rl q- r2)
!
!
With the definitions Avl = v I - Vl a n d A V 2 - - V2 - - v 2 , the magnitudes of the velocity i n c r e m e n t s that m u s t be delivered by the two thrusts are AVl-- V/ v2 + v'I2 - 2 V l v'1 cosfll,
AV2-v/Va2+ v~a2-2v2v~acos[32
(5.28)
as follows from the cosine law applied to the triangles f o r m e d by the velocities. The transit time t2 - ta from the initial to the final orbit is one-half of the period of the transfer ellipse, h e n c e h
t~ = 7v
2
(5.29)
I m p o r t a n t for estimating the total p r o p u l s i o n r e q u i r e m e n t for the maneuver is the s u m A v = AVl + A v2. It is easily shown from (5.27) a n d (5.28) that in the case of no plane change and for a fixed initial orbit radius, Av at first increases with increasing radius r2, t h e n decreases again. The m a x i m u m required Av occurs at a radius ratio r2/rl of 15.58, where AV/Vl = 0.536. In the limit, as r2 / rl ---> r A v/Vl " ' > ~ - - 1. In the more often e n c o u n t e r e d case i n which t h e r e is a plane change, the s u m of the angles/~1 a n d f12 will be known. The split b e t w e e n the plane change at periapsis and the one at apoapsis is o p e n to the mission designer. Usually, the a p p r o a c h used is to minimize the s u m Av of A Vl a n d A v2.
C H A P T E R 5 OrbitalManeuvers
194
Since vlv~/(vav'2) - (re/r1) 3/2, the factor of cos fll in (5.28) is larger (for r2 > rl) than the corresponding term ,
/z /
2rl
v2v2 = -~2V rl + r2 For instance, for a geostationary orbit, to be reached from a low earth orbit, the ratio (ra/rl) 3/2 is about 17. Reducing/~1 at the expense of/~p will reduce Av and therefore also the propellant consumption. The optimal split between the two plane changes will depend on the radius ratio r2/rl. In the important case of the geostationary final orbit, starting from a low-altitude parking orbit, the split will be optimal if almost the entire plane change is done at the apogee. The benefit derived from introducing a partial plane change also at perigee (where the velocity is higher) is no longer significant. The H o h m a n n transfer orbit satisfies the condition that the second (empty) focal point of the ellipse lies on the line connecting the transfer's initial and final points. It then follows from the discussion in Sect. 5.2 that the H o h m a n n transfer is the m i n i m u m energy path between the Hohmann periapsis and apoapsis.
5.5
Other Transfer Trajectories Transfer trajectories other than H o h m a n n transfers are frequently used. The choice is often dictated by the need to match the impulse of solid-propellant motors to the mission requirement. These motors must burn to completion and may have excess impulse for the intended spacecraft mass. (Sometimes, the motors are "offloaded," that is, filled with less propellant than the capacity of the motor case would allow; offloading, however, requires a new mandrel for the casting process.) The type of transfer orbit that is optimal will depend significantly on the radius ratio or semimajor axis ratio of the final orbit relative to the initial one. For transfers from an initial circular orbit to a final circular orbit and a radius ratio larger than about 15, it may happen that the sum of the A v's for n o n - H o h m a n n transfers is slightly less, at best about 4%, than for the H o h m a n n transfer. (Because the point at which the transfer path meets the final orbit is allowed to vary, this does not contradict the earlier statement that the H o h m a n n trajectory is the path of least energy between the H o h m a n n periapsis and apoapsis points.) Often it is required not only that a final orbit be achieved but also that the time phasingbe correct, that is, that the spacecraft arrive in its final orbit at the required place and time. In particular, this is the case in rendezvous maneuvers where the spacecraft must meet, and possibly dock, with another spacecraft. The transfer time will generally be longer than would be the case in the absence of such a requirement. In what follows in the present section, it will be assumed that the initial and final orbits are circular and that their planes a n d also the transfer plane are coplanar. In Fig. 5.9 three different types of transfer orbits
5.5 Other Transfer Trajectories
195
V2
BE
I
/
V2 x
x
~V
~ v2
J/ I I I I I I I I I
/
,
,
I
HM
I I
\ \
,.,-/
\
V 2 \
/
/
/
BE
ST
/
/
/
2"
\
\
i//v2
\
I I
v3
9= Rocket M o t o r Burn v2
V2
Figure 5.9 Comparison of four transfer trajectories: H, Hohmann; HM, Hohmann modified; BE, bielliptic; ST, semitangential. (modified Hohmann, bielliptic, and semitangential), all for the same initial and final orbits, are shown and compared with the H o h m a n n transfer. For these orbits, such data as the velocity increments and the time needed for the transfer can all b e calculated by using the algebraic equations derived in Chap. 3. However, transfers between elliptic orbits can be calculated in this way only in some special cases. In all other cases, recourse must be made to numerical methods. The modified H o h m a n n transfer is a two-impulse path similar to the Hohmann transfer but has a somewhat higher apoapsis. As the transfer orbit meets the final orbit, a small change of the flight angle, in addition to the velocity increment, is made. This maneuver is frequently advantageous when solid-propellant motors are used. The basic requirement is that the propellant be fully c o n s u m e d when the spacecraft has the correct, or very nearly correct, velocity for the desired final orbit. To avoid needless repetition, we indicate simply the sequence of calculations by their equation numbers in Chap. 3. Given/z, r~, 1"2,and an assumed A Vl, the sequence used to find A v2 and the transfer time is as follows: From/z and r~ follows v~ Adding to this A v~ gives v~ and the angular m o m e n t u m h' of the transfer path. From (3.10), with 0 = 0, follows the transfer paths eccentricity e'. From (3.11) follows the energy, w', and hence from (3.12) the semimajor axis. From 1"2and (3.10) follows the true anomaly 02 at the (second) intersection of the transfer path with the final orbit. The
196
C H A P T E R 5 OrbitalManeuvers radial and azimuthal c o m p o n e n t s of v~ follow from (3.14). The required zkv2 = v2 - v~, needed to start the final orbit, is then immediate. Finally, the eccentric anomaly E~ at 82 is found from (3.20), and the transfer time, t2 - tl, is calculated from (3.21). The bielliptic transfer, for instance, between two circular, coplanar orbits (Fig. 5.9), uses three impulsive thrusts. The path consists of two semiellipses, the first one tangential to the initial orbit, the second one to the final orbit. There are no changes in flight angle. Prior to orbit insertion, the spacecraft reaches an altitude higher than the altitude of the final orbit. At the point of termination of the second semiellipse, thrust is applied tangentially to the path but in the reverse direction ("retro-thrust"). The calculation can proceed in the same m a n n e r as just outlined. The s e m i t a n g e n t i a l transfer is qualitatively the same as the modified H o h m a n n transfer but usually goes to a higher altitude. Depending on the choice of the semimajor axis, a favorable time phasing, important for instance for some rendezvous maneuvers, can be obtained. Insertion can take place at either point I or ], as may be closest for matching the rendezvous requirement.
5.6
On-Orbit Drift Once a spacecraft is in its intended orbit, a relatively modest expenditure of propellant will suffice to change the time phasing of the spacecraft. The necessary m a n e u v e r consists of two short burns of a low-thrust motor. The thrusts are nearly tangential to the orbit. The first thrust imparts a small incremental velocity, either in the direction of the orbital motion or opposite to it. The result is a relative drift motion, which is stopped by a second thrust in the direction opposite to the first. This m a n e u v e r makes it possible, for instance, to change the longitude of a geostationary satellite. If several satellites, with one or more perhaps serving as spares, have been placed in the same orbit, their relative positions can be changed and a satellite made to assume the function of another, perhaps disabled, one. There are still other applications, for instance, in rendezvous maneuvers. Prior to docking, when the two spacecraft are already on the same orbit, although possibly still widely separated, the drift velocity imparted to one of them will, over time, close the distance. Prior to final orbit insertion, or before the start of a hyperbolic trajectory, spacecraft are very often put into a parking orbit about the earth. The usual purpose is to check out the spacecraft functions, as they may have been affected by the launch. The choice of the parking orbit and time phasing is dictated by the communications requirements imposed by the location of the ground station that does the checkout. After completion of the checks, which m a y take several weeks, the on-orbit drift m a n e u v e r can be used to place the spacecraft into its operational position and phasing or into a new position for further acceleration. Assuming for simplicity a circular orbit, the a m o u n t of propellant consumed is easily estimated. (The slight increase in eccentricity that follows
5.7 Launch Windows
197
the first thrust can be neglected for this purpose.) If r is the radius of the orbit, ms the mass of the spacecraft, I~p and rh the specific impulse and mass flow rate of the motor, and Atpr the duration of the burn, the impulse given to the spacecraft is go Isp m A tpr. Hence the drift velocity, A v, imparted by the first thrust is
Av--
go lsp A mpr ms
where Ampris the mass of propellant c o n s u m e d by each of the two thrusts. If the drift is to advance (or retard) the spacecraft by a polar angle (true anomaly) difference A0 in comparison with the u n p e r t u r b e d motion, and if At is the time allowed for the drift, the a m o u n t of propellant c o n s u m e d by the two thrusts is 2Ampr =
2AOrtas goIsp At
(5.30)
Taking as an example a geostationary spacecraft (r - 42 164 km), ms = 1000 kg, Isp = 200 s, A0 -- 90 ~ and At -- 60 days, the total propellant consumption calculated from this equation is 2Ampr -- 13.0 kg.
5.7
Launch Windows The final orbit plane of a satellite, or the plane of the trajectory of a spacecraft on a deep-space mission, is determined by the plane's inclination to the equatorial plane and by the right ascension of the ascending node (Sect. 3.4). Because of the need to place the spacecraft trajectory in the selected plane and taking into account the earth's or planetary rotation and with it the motion of the launch site ~ the acceptable launch times are restricted. Also, because of various operational requirements, margins in the time of launch must be provided for. Providing these margins requires that the flight and ground computer software be designed to initiate the necessary adjustments of the flight path whenever the actual launch time differs from the nominal one. Time intervals acceptable for a launch are referred to as l a u n c h windows . Other things being equal, the launch window for a launch from the earth repeats itself approximately every 24 h. Depending largely on the required inclination of the final orbit, launch windows may vary in duration from a few minutes to several hours. As far as astrodynamic considerations are concerned, the time of launch into equatorial orbits, in principle, is not restricted. On the other hand, near-polar orbits necessitate very short windows. Otherwise, excessive propellant c o n s u m p t i o n might result from the plane changes that would be needed if the actual and nominal times of launch differed by more than a few minutes. In addition to the requirement of placing the spacecraft into the selected orbital plane, it may be necessary to control the time at which the spacecraft passes its ascending node, narrowing the available window further.
198
C H A PT E R 5
OrbitalManeuvers
Launch windows can be widened if there is excess propellant either in the launch vehicle or in the upper stages. Liquid-propellant systems always carry a small a m o u n t ~ sometimes as little as 1% ~ of excess propellant to cope with uncertainties, for instance, in the propellant mixing ratio or in the initial propellant loading. If the nominal flight path and the available motors are such that excess propellant above the a m o u n t needed for a margin of safety can be carried, the launch window can be widened. The excess propellant then becomes available for plane changes and other flight path adjustments. Other than the restrictions imposed by astrodynamics, operational considerations also influence the choice of an acceptable time of launch. Such operational aspects include the weather at the launch site, particularly highaltitude winds, the sun angle on the spacecraft before and after deployment of solar panels, the need sometimes for sufficient daylight visibility of the launch, limitations caused by the boil-off of cryogenic propellants, and the availability of all of the needed communications links. Figure 5.10 is a graphic representation of a launch window for a particular mission. It applies to the placement of a payload into orbit by means of the U.S. Space Shuttle and two upper stages launched from Cape Canaveral. The Space Shuttle m e a n altitude was 278 km. The final orbit of the payload was circular, at an altitude of 20,187 km (a 12-h orbit), with an inclination to the equator of 55.0 ~. The two upper stages released by the Space Shuttle provided two additional velocity increments of 2107 m / s by the lower and 1888 m / s by the upper stage. Both solid-propellant motors of these
Figure 5.10
Launch window for a Cape Canaveral Space Shuttle launch for a mission with two upper stage firings. N-N, maximum inclination allowed by range safety. Adapted from Chobotov, V. A., "Orbital Mechnics," 2nd ed., Ref. 4.
5.8 Injection Errors and Their Corrections
199
upper stages had excess propellant. The chosen flight path therefore was a modification, by raising the apogee and a plane change, of a H o h m a n n transfer. In the figure, the ordinate represents possible inclinations of the flight path of the Space Shuttle when launched from Cape Canaveral. The right ascension of the Space Shuttle's path is the abscissa in the figure and is shown as the difference from the right ascension of the payload orbit. The abscissa therefore also represents the delay, or advance, of the true launch time (e.g., a 15 ~ increase of the right ascension corresponds to a delay of 1 hour). Also represented in the diagram is the payload orbit. The shaded area represents the Space Shuttle's launch window that is compatible with the payload's final orbit and with the a m o u n t of propellant in the upper stage motors. If it is launched at the nominal time, the launch inclination, because of the excess propellant available in the upper stages, could be anywhere from about 28 ~ to 41 ~. The window's boundaries are limited, however, by the m a x i m u m launch inclination (about 57 ~ corresponding to a launch azimuth of 35 ~ north) allowed by range safety considerations at this particular launch site.
5.8
Injection Errors and Their Corrections The injection for instance, by a H o h m a n n transfer into an intended orbit will unavoidably result in some random errors in the orbital parameters. These must be corrected if the high precision that is usually required for the final orbit is to be met. The corrections are typically made by first reorienting the spacecraft in the direction needed for the thrust, followed by a short burn of an upper stage motor. The following discussion is limited to the case of final orbits that are circular. The extension to elliptic orbits is straightforward, however. If the final, intended orbit is circular, the approximate orbit, that is, the orbit resulting from a H o h m a n n or similar transfer, can have at most three sources of error. They are an error, designated by Ai, of the inclination of the orbital plane; an error, designated by Aa, of the semimajor axis; and an error, designated by Ae, of the eccentricity. It is convenient to discuss these error sources separately. The total error and the required corrective maneuvers can then be obtained by superposition.
5.8.1
Inclination Error
Figure 5.11 shows the geometrical relations represented on the unit sphere. Using primed symbols for the approximate and u n p r i m e d ones for the final orbit, there are A i= i'-
i,
a ' = a,
e'-- e = 0
where [Ail << 1 by assumption. Let AS(t) be the spacecraft's latitude and ~'(t) the spacecraft's longitude before the correction. Both latitude and longitude here refer to the final orbital plane. The origin oftime is taken at the m o m e n t
200
C H A PT E R 5
Orbital M a n e u v e r s
Figure 5.11 Orbit plane inclination error and correction. w h e n the spacecraft on the approximate orbit crosses the final orbital plane. Since the semimajor axes (here the radii) are, by assumption, the same for both paths, the "mean angular velocities" (Sect. 3.2)
/It=/1= ~a~3 are also the same. From the law of sines applied to the spherical, right angle triangle shown in the figure, sin A~ = sin Ai sin(m) Expanding sin A~ and sin Ai into their power series and omitting third-order terms, zX~ = A i s i n ( n t ) ,
(a'=a,
e'=e=O)
(5.31)
The spacecraft on the approximate orbit is therefore seen to change its latitude relative to the final orbit sinusoidally in time, with amplitude Ai. Let the longitude difference A~ = ~.' - ~., where )~ = nt. Then, from the same spherical triangle, cos(nt) = cos )~' cos A~ = [cos(nt) cos A)~ - sin(nt) sin A)~] cos A~ Expanding in power series the trigonometric functions of the perturbation terms,
1 1 cos(m) = [ c o s ( n t ) ( 1 - ~A)~2 + .. . ) - sin(nt)(A)~- ~A)~3 + ... )] x ( 1 - 21 A ~ 2 + . . . ) After dropping the higher order terms and substituting A8 from (5.31), the final result for A~. to the lowest significant order becomes A)~ -- _1~ s i n ( 2 n t ) A i 2,
(a' = a, e' -- e = 0)
(5.32)
5.8 Injection Errors and Their Corrections North f
1.0 ~
nt =
1
3
/ -
G r o u n d Track prior to Insertion
/
- Final Position
/
0.5 ~
A5
201
0o
0,
/
-0.5 ~
2
West
-1.0 ~ -0.01~
o
Z
0o
0.005 ~ 0.01~
East
Ak
Figure 5.12 Ground track of geostationary satellite prior to correction of 1~ inclination error. The longitude difference is therefore only of s e c o n d order in Ai, with a frequency twice that of the latitude difference. The error in inclination can be corrected by applying thrust p e r p e n d i c ular to the plane o f m o t i o n , at either p o i n t I or point J in the diagram, w h e r e the two planes intersect. The n e e d e d velocity i n c r e m e n t A v is seen to be Av = Ai v, where the velocity v is from (3.14), so that Av = v/--~/a A i
(a' = a, e' -- e -
0)
(5.33)
The geographic latitude a n d longitude are o b t a i n e d from A8 a n d Ak by a straightforward coordinate transformation. A particularly simple case occurs for geostationary satellites. Shown in Fig. 5.12 is the g r o u n d track, that is, the radial projection o f t h e g e o s t a t i o n a r y satellite's p a t h onto the earth surface. An inclination Ai of 1.0 ~ is a s s u m e d . The m e a n angular velocity in this case is n = 2n/dsi (dsi = length of sidereal day). As given by (5.31) a n d (5.32), the satellite in its a p p r o x i m a t e orbit before the corrective m a n e u v e r is seen to describes a figure eight, with a m u c h smaller variation in longitude t h a n in latitude.
5.8.2
Semimajor Axis Error Let Aa -- a' -- a and An-- n'- n=
(a + Aa) a -
Expanding by m e a n s of the b i n o m i a l t h e o r e m a n d retaining only the lowest significant order terms in A a / a gives, An=
sa
(Ai = 0, e ' - - e = O)
(5.34)
C HA P T E R 5
202
OrbitalManeuvers
This shows that for a positive (negative) Aa, prior to the satellite's injection into the final orbit, there is a slow westward (eastward) drift relative to the final motion. The corrective maneuver consists of a H o h m a n n transfer between the two coplanar circular orbits (Sect. 5.4). Therefore two impulsive thrusts are needed. The two required velocity changes A Vl and A v2, obtained from (5.27), are, to lowest order in Aa/a,
AVl= / Av2 =
/z
a+Aa-
~
2/za
(a+Aa)(2a+Aa)
~/2/z(a+Aa) ~ a(2a+ Aa) -
=4
l ~ a~
Aa
(5.35a)
l ~ a~ = ~ Aa
(5.35b)
(Ai = 0, e' = e = 0)
(5.35r
hence Avl = Av2,
5.8.3
Eccentricity Error Let Ae = e' - e = e' since, by assumption, e = 0. The eccentric anomaly, E', for the orbit prior to the correction is, from the Fourier-Bessel series (3.30), 1 -klk(k Ae) sin(kM')
E ' = M' + 2 ~
k=l where M' = n' (t - tp) = n(t - tp) since the m e a n angular velocity, n, depends only on the semimajor axis. The time of periapsis passage is designated by tp. Expressing the Bessel functions by their Taylor series and dropping terms of third order in Ae and higher, the final result, expressed by the difference AE = E ' - E between the eccentric anomalies before and after the corrective maneuver, is therefore A E = A e s i n ( n ( t - tp)) +
1( A e ) 2
sin(2n(t - tp))
(Ai -- 0, a' -- a)
(5.36)
In place of the difference of the eccentric anomalies, it may be more convenient in some applications to compute the difference of the true anomalies. They, as well as the radial difference, can be obtained, for instance, from (3.20). The corrective maneuver can take place at either the periapsis or apoapsis. Computing the needed velocity change is straightforward.
5.9
On-Orbit Phase Changes It is sometimes necessary to change the position of a spacecraft on its orbit without changing the orbit itself. What is being altered, therefore, is the time p h a s i n g of the spacecraft motion. For instance, because of changing patterns in communications traffic, there may be a need to move a geostationary communications satellite along its orbit to a different longitude. Another such instance may occur when a
5.9 On-Orbit Phase Changes +
rO o
203
O_
/ ! ! !
i
T-
I \ \ \
"
~....._
..~ ~
-"
0+
1
1
x Spacecraft Positions with or w i t h o u t Maneuver
Figure 5.13 Changeofphaseofa spacecraft on a circular orbit. 00, basic orbit; O+ and 0_, transfer orbits; C, center of attraction; T, location of thrust. constellation of satellites, all on the same orbit, needs to be rearranged because one of the spacecraft has reached the end of its useful life. In what follows, we assume, for simplicity, a circular orbit, designated by O0 in Fig. 5.13. The extension to elliptic orbits is straightforward, however. To accomplish the change, two impulsive thrusts, both at point T in the figure, are applied, either in the direction of motion or opposite to it. The two thrusts are assumed to be separated by a time difference A t. If the first thrust is in the direction of motion, the spacecraft will follow an elliptic transfer orbit O+ which has an increased semimajor axis, hence longer period. This orbit will lead the spacecraft back to point T. After one or, in more c o m m o n applications, m a n y revolutions, a second, short thrust in the opposite direction will place the satellite back on its original orbit. Correspondingly, if the first thrust is in the retro-direction, the spacecraft will at first follow the transfer orbit O_, before a second thrust will bring it back to O0. The point T is the periapsis of O+ and the apoapsis of O_. The result of following O+, with its longer period, will be a satellite position behind the one without the maneuver. Correspondingly, by following O_, with its shorter period, the satellite will be ahead. To minimize the expenditure of propellant, the two impulses are kept small, resulting in transfer orbits that differ little from the original orbit. Unavoidably, the time needed for completion of the m a n e u v e r then becomes long, sometimes as m u c h as several months. Let a0 be the radius of the circular orbit O0 and al the semimajor radius of a transfer orbit. It follows from the geometry of ellipses that ao = a l ( 1 7: el)
(5.37)
where el is the eccentricity of a transfer orbit and the u p p e r sign here and in what follows applies to orbits of the type 0+, the lower sign to orbits 0_.
204
C HA PT ER 5
OrbitalManeuuers
The periods P0 of the original (circular) orbit and P1 of the transfer orbits are to lowest significant order
Po -- 2zr~ -aa~ -,
P1 -- 2 J r , / l (
/~
a0 1Tel
)3
-- Po(1 T e~) -a/2
(5.38)
If nl is the n u m b e r of revolutions on a transfer orbit, the time between thrusts, hence the time needed to execute the maneuver, is At = n~/)1. In this time, the spacecraft moves through a polar angle (true anomaly) of 2zrAUP1. Without the maneuver, the angle would be 2zr At/Po. Therefore the difference
(1 1)
A4) -- 2JrAt /)1
Po
is the ultimate angular position of the satellite after carrying out the maneuver relative to what the position would have been without the maneuver. For transfer orbits illustrated by O+ in the figure A~ is negative; that is, the spacecraft is delayed by the maneuver. Correspondingly, for transfer orbits illustrated by O_, the spacecraft is advanced. Substituting for the periods from (5.38), expanding by the binomial series, and retaining only the lowest order terms in el result in the approximation A ~ --
T-~
elAt
(5.39)
Let v0 be the velocity of the spacecraft before the first and again after the second thrust. At the periapsis of O+ (apoapsis of O_) the velocity Vl, from (3.14) and (5.37), is
Vl -
j(2 #
ao
1) __ ~a~0(1 + el)
al
so that A v~, the change in velocity by the first thrust, and A re, by the second thrust, are AVl -- - A v 2
-- Vl - v o - -
( 1 -1- el) - V / ~ o~
1 / ~----el
(5.40)
= +2V a0
where again the binomial expansion for small el was used. Combining (5.39) and (5.40), ].
A v l A t - - - 5aoA4)
(5.41)
valid for either orbit. This shows that for a given displacement Ar the quantity 21Avl = 2lAv~l that characterizes the expenditure of propellant by the maneuver can be made arbitrarily small, although only at the cost of a
5.10 Rendezvous Maneuvers
205
longer delay time At. For advancing (retarding) the spacecraft by an angle between 0 and 180 ~ a smaller 21A v I will result w h e n the thrusts occur at the apoapsis (periapsis) of the transfer orbit. For instance, if it is necessary to increase the geographical longitude (i.e., a displacement toward the east, hence in the (assumed) direction of the motion of the satellite) by an angle less than 180 ~ a velocity decrease by the first thrust and a velocity increase by the second thrust are required. The reverse will apply for a displacement toward the west. To consider another example: A navigational system may have eight equally spaced satellites on each of its (circular) orbits. Replacement of one of the satellites by one of its neighbors will therefore require an angular displacement of 45 ~ For a radius a0 = 20,000 km and allowing a time interval At of 30 days, IAvll = lava1 from (5.41) becomes 2.02 m/s. Because of the lack of exact sphericity of the earth, geostationary satellites tend to drift slowly from their intended station. The drift in the n o r t h south direction resulting from the 12 term in the gravitational potential (Sect. 2.1) is considerably larger than the drift in the east-west direction. C o m m o n practice is to correct the drifts by firing small thrusters on the spacecraft, perhaps every few days or every day. The need for frequent n o r t h - s o u t h corrections can be reduced if the plane of the orbit is allowed to have a small inclination relative to the equatorial plane. This will save satellite propellant. For c o m m u n i c a t i o n satellites, a complication, however, is that spot b e a m s intended for downlinks to small receiving areas need to be steered, either mechanically or electronically.
5.10
Rendezvous Maneuvers For the purposes of a rendezvous and, possibly, docking of two space vehicles, not only their orbits but also their positions on the orbit at a time must be matched. For docking, the relative angular orientation ("attitude") of the two vehicles must also be controlled. In the present section, we consider only the motions, and particularly the relative motions, of the centers of gravity of the two vehicles. Rendezvous of spacecraft are useful for such purposes as the resupply of a spacecraft or space station by a cargo vehicle, for moving a spacecraft to a new orbit by m e a n s of a "space tug," or for a vehicle ascending from a planetary surface to meet an orbiting vehicle. It is useful to distinguish several phases in the rendezvous maneuver. The ones that we consider here are 1. The transfer of one of the vehicles, here called the "active vehicle," from a parking orbit to the approximate orbit and time phasing of the second, or "passive vehicle." This phase of the motion is formulated in the geocentric (or planetocentric or heliocentric, depending on the application) reference flame, which for the present purposes is a very close approximation of a true inertial flame.
206
C HA PT E R 5
OrbitalManeuvers
2. The closing m a n e u v e r that brings the two vehicles together a n d starts from what was at the end of the first phase only an approximate vicinity. To achieve the required precision, this phase is formulated in a reference flame that is fixed in the active vehicle. In general, this will not be an inertial flame.
5.10.1
Rendezvous Maneuver, First Phase As illustrated in Fig. 5.14, it is a s s u m e d that the active spacecraft is already on a parking orbit, which is taken here to be circular. For simplicity, the passive vehicle is a s s u m e d to be on a coplanar, also circular, orbit. The two orbits are c o n n e c t e d by a H o h m a n n transfer semiellipse. The maneuver therefore requires two impulsive thrusts by the active vehicle, one in leaving the parking orbit and a second one at the end of the transfer so that the positions, speeds, a n d directions of m o t i o n of the two vehicles can be matched. Let rl and r2 (r2 > rl) be the radii of the two orbits. The corresponding periods are, from (3.16), P1 = 2yr/z-l/2 /...3/2 1 ,
P2 = 27r #-1/2..3/2 12
(5.42)
The semimajor axis o f t h e H o h m a n n transfer orbit is (1/2) (rl q- r2). If tl is the time of the initiation of the H o h m a n n transfer by the active vehicle, t2 the time of the rendezvous, then
t2-tl = ~ #
-1/2(r1+r2) 3/22 (5.43)
--2-S/2p1(1+r2)
RV
t2
t~ to
Figure 5.14 Rendezvous maneuver, first phase, with Hohmann transfer. Circular orbits with periods P1 (active spacecraft) and P2 (passive spacecraft). RV, rendezvous point; C, center of attraction.
5.10 R e n d e z v o u s M a n e u v e r s
207
If at some arbitrary initial time t = to, the angular separation of the two vehicles is A | where A|
-- 27rn + A0
where 0 _< A0 < 27r. Here, n - 0, 1, 2, ... indicates the n u m b e r of additional full revolutions on the circular orbit that the active vehicle m a y have to m a k e to m e e t the rendezvous condition. The angular separation (true a n o m a l y difference) A0 b e t w e e n the two vehicles on their circular orbits is r e p e a t e d regularly after each interval of time, Psn, called the s y n o p t i c period. (The term is borrowed from astronomy, where it designates the time interval from one approximate a l i g n m e n t of sun, earth, and a planet to the next such alignment.) It follows from equating the polar angles of the two vehicles that
27r Psn / P1 -- 27r (Psn / P2 q- 1) therefore
Psn -- P1 P2 / (P2 - P1 )
(5.44)
The condition for the two vehicles to m e e t at the same p o i n t in space a n d time therefore b e c o m e s 2zr (tl - to) 27r ( t 2 - to) +r~-+AO P1 P2 or
PIP2(AO/27r-1)+zrPl/~-l/2(rl-+-r2) 3/22 () 2-5/2p1((r2/rx)+l)3/2
(P2- P1)(tl- to)from which tl - tO --
AO 2zr
1 Psn -4-
2
(r2/rx
)3/2
-1
(5.45)
from (5.42) and (5.43). This time difference is the n e e d e d delay before the active vehicle initiates the H o h m a n n transfer. To illustrate the result by n u m e r i c a l data: For a parking orbit at 300 k m altitude above the earth and a geostationary satellite, the two periods P1 a n d P2 are 90.5 and 1436 minutes, respectively. The synoptic period, Psn, from (5.44), is 96.6 minutes. The time t2 - tl n e e d e d for the H o h m a n n transfer, from (5.43), is 316.5 minutes. If 6) is a s s u m e d to be 180 ~ the delay time tl - to before initiating the transfer orbit, from (5.45), is 21.3 minutes.
5.10.2
Rendezvous Maneuver, Second Phase The first phase of the m a n e u v e r will bring the active vehicle into the vicinity of the passive vehicle. But the reference frame a n d the two separate Kepler orbits for the two vehicles that have b e e n used in the preceding section are n o t accurate e n o u g h to formulate the close-in final a p p r o a c h of the two vehicles. Instead, a suitable reference frame for describing this s e c o n d p h a s e is one that is fixed in the a c t i v e vehicle. In general, this will n o t be an inertial frame. This frame also has the advantage of being directly applicable to the radar and optical sensors on the active vehicle. These sensors image the
208
C H A PT E R 5
Orbital Maneuuers CO
x2
Active Vehic
I
[ /
/
~
/ / s
]/~--fX1
Passive-/ Vehicle
/
~- Center of Attraction
Figure 5.15 Docking maneuver: illustration of Eq. (5.47). passive vehicle and are n e e d e d in the p e r f o r m a n c e of a close-in m a n e u v e r to the necessary precision. In what follows we will be c o n c e r n e d only with the relative m o t i o n of the centers of mass of the vehicles, not with their relative angular orientation, which, of course, m u s t also be closely controlled before docking. Referring to Fig. 5.15, Xi, i - 1, 2, 3, are the Cartesian coordinate axes of the (nonrotating) frame of reference with origin in the mass center of the planet. Fixed to the active vehicle, with the origin in its center of mass, the xi axes define the second reference frame that will be used. We designate the Xi c o m p o n e n t s of the radius vector R from the center of the planet to the center of mass of the active vehicle by Ri, and correspondingly the c o m p o n e n t s of the radius S to the passive vehicle by Si. The components of the vector r = S - R in the xi reference frame are designated by ri. The angular velocity (relative to an inertial reference) of the active vehicle is designated by w, with c o m p o n e n t s ~oi in the xi reference frame. We designate by m the mass, by Fg the force of gravity, and by Ft the thrust, all referring to the active vehicle. Time differentiations are indicated by d( ) / d t in the Xi reference frame a n d by (') in the xi frame. In indicial notation, the equation of m o t i o n of the active vehicle in the planetocentric reference frame Xi, from (2.1), is d 2 Ri = dt 2
# Ri Ft, i + ~ R3 m
(i - 1, 2, 3)
(5.46)
To express the equation of m o t i o n of the passiue vehicle in the frame xi that is fixed in the active vehicle, we m a k e use of the symbol
eijk
- 1 -------1 - 0
if (i, j, k) is an even p e r m u t a t i o n of (1, 2, 3) if (i, j, k) is an odd p e r m u t a t i o n of (1, 2, 3) otherwise (i.e., the same index appears at least twice)
The Cartesian c o m p o n e n t s of vector products, such as the ith c o m p o n e n t of a x b are therefore (a x b ) i - E eijkajbk j,k
In this n o t a t i o n the e q u a t i o n (1.4) for transformations b e t w e e n two reference frames in relative m o t i o n to each other, w h e n applied to the passive
5.11
Gravity Turn
209
vehicle, becomes ll, Si
..
ri =
d 2 Ri
dt2 +~j (~o~ri- ~oioojrj) - ~j,k eijk(2~oji'k +rkbj)
S3
The first term on the right can be expanded by the binomial series. Retaining only first-order terms in [ri/R[ << 1, 3 2 --
Z ( R j + rj) 2 -- ~ Rj 1 + --~j} J J
1 + ~ ~ Rjrj j
Using the binomial expansion once more results in
~Si- -
Ri(l +
R (I+- ZjRjrj)
3/2
~-,~ Rj ]
= R3 R i 1[ rqi 3
~
"
rj
Combining this with (5.46), the zero-order terms are seen to cancel so that only first-order terms in rg and its derivatives remain. When solved for the c o m p o n e n t s of the control thrust Ft, the final result obtained is Ft, i m =
Ri [ ri r'i--lZ~
Ri
3 R 2 Z. R j r j J
]
+~ (riw~ ooiwjrj) y~. eijk(2wji'k +rk&j) j
j,k
( i - 1, 2, 3) (5.47)
This equation is basic to the development of the software for computing the thrust that is needed to control the active vehicle on a prescribed approach path. The vehicle's angular velocity w and its derivative are obtained from inertial m e a s u r e m e n t units or similar devices based on the use of mechanical or ring-laser gyroscopes. At least three gyros and three rate gyros are needed. The distance vector r and the range rate are c o m p u t e d from data that are provided by radar and optical sensors. In this m a n n e r the required thrust as a function of time can be obtained by a closed-loop feedback control that compares at each instant the desired and actual distance vectors and initiates the needed corrective thrust.
5.11
Gravity Turn Large launch vehicles normally take offfrom the launch site nearly vertically. Similarly, space probes, after having landed on a planet or moon, take off along a vertical. To make the transition to an orbit, the trajectory must then turn away from the vertical. Because of their stringent weight limitations, launch vehicles cannot be designed structurally to withstand appreciable aerodynamic forces and moments. For this reason, the attitude of the vehicle, while in the denser part of the atmosphere, needs to be controlled to keep the angle of attack within at most a few degrees. Upper atmosphere winds and wind shear aggravate the structural and steering problems.
210
C H A PT E R 5
Orbital M a n e u v e r s V
C
84
Trajectory
Figure 5.16 Gravity turn, schematic. C, center of mass of launch vehicle. A trajectory that turns from a near-vertical path, yet minimizes the angle of attack, is call a gravity t u r n [5]. Once a slight deviation from the vertical (the "kick angle") has been initiated by steering, gravity alone will deflect the trajectory from the near vertical toward a more horizontal flight p a t h (Figs. 5.16 and 5.17). The turning of the trajectory will generally take place over a distance small c o m p a r e d with the radius of the earth or planet. Therefore the curvature of the planet's surface can be neglected, and so can the change with altitude of the gravitational acceleration. As indicated in Fig. 5.16, let x(t) be the horizontal, z(t) the vertical position of the center of mass of the launch vehicle, v(t) the velocity, and y (t) the flight angle. Ft is the thrust, Fd the drag force acting on the vehicle. Both forces are nominally parallel to v. The m a g n i t u d e of the thrust will be a s s u m e d to be constant, whereas the drag will change with altitude and velocity. We write F - Ft + Fd. In coordinates fixed to the planet,
-.20
T
t33
<
re(t) d v x / d t -
F(t)vx/V
re(t) d v z / d t -
F(t)vz/v - gin(t)
-.16
-.12
X"
-.O8 -0.10
/
-0.30
/-1-o4o
(5.48)
-.O4
0 t .I = 0 /
-0.02
-0.04
/---0.5o
-0.06
11-0.60 ----0.70
-0.08 245 km (for/sp = 350 s)
Figure 5.17 Solution of the equations (5.55) for a gravity turn. For g = go, v' (t' = 0) = 1.00, k = 4.00. Drag neglected.
5.11
Gravity Turn
211
where m(t) is the vehicle's mass and g - const, the gravitational acceleration. Multiplying the first equation by vz, the second by Vx, and subtracting v~ d v x / d t
-
(5.49)
Vx dye~dr - gvx
From d
~(tang)-~
1(
d(vz) ~
- -v~
dvx dvz) v~--~ - Vx-~ -
g Vx
follows dF dt
g = - - cos y
(5.50)
v
The temporal derivative of the speed is given by dv
=
dt
F
g sin ~,
m
(5.51 )
the change in altitude by dz dt
= v sin y
(5.52)
It is convenient to use ~, a n d v as the d e p e n d e n t variables. In addition to this (nonlinear) system of differential equations, there are the auxiliary relations
F~
m(t) - mo
gOIsp
(5.53)
t
and
(s.s4)
Fd -- A C d ( M ) ~ p1 a ( Z ) v 2
The former follows from (2.24) and Ft -
go l s p m -
-g0Isp d m / d t
Here, rh is the propellant mass flow rate, a s s u m e d to be constant, m0 the vehicle mass at t = 0, A the vehicle m a x i m u m cross-sectional area, Cd(M) the drag coefficient, and Pa (z) the a t m o s p h e r i c density. The drag coefficient d e p e n d s principally on the Mach number, M. The solution of the system of equations (5.50) to (5.54) is obtained numerically. Neglecting the a e r o d y n a m i c drag can serve as a useful first approximation. An example is shown in Fig. 5.17. It is convenient in this case to introduce the n o n d i m e n s i o n a l quantities
t'-- t/Isp, x' = x/(golasp),
m ' - m / mo ,
v'
v / (go lsp ) ,
k-
Ft/(gomo)
Y'= Y/(goG)
where k is a constant. If, as a first approximation, a e r o d y n a m i c drag is neglected, the equations of m o t i o n in terms of the n o n d i m e n s i o n a l quantities
212
C H A PT E R 5
Orbital Maneuvers
become dy
g cos y
dr'
go
dr' dr'
l)t
k 1
-
(5.55) g sinF
kt'
go
To avoid the singularity that arises when the flight angle approaches 90 ~, it is advantageous to integrate the equations backward in time, starting, for instance, at the point where the flight path would become horizontal. An example of the numerical solution of these equations is shown in Fig. 5.17. The calculation applies to a launch from the surface of the earth (g = go). The origin of time is taken at the point where the flight path would have become horizontal. The parameters chosen are v'(t' = 0) -- 1.00, k -4.00. For an assumed Isp of 350 s, the horizontal distance traveled from the point of launch to the point of zero flight angle becomes 245 km. The velocity at this latter point becomes 3430 m/s.
Nomenclature a e
E F
f,g g i m t/
P r,R S v t/)
AW Y 0 # ~o
()m ()p ()pr
semimajor axis eccentricity eccentric anomaly force variables in Lambert's theorem, Eqs. (5.9) gravitational acceleration; go: standard gravitational acceleration inclination (Fig. 5.4) specific impulse mass m e a n angular velocity; n u m b e r of complete revolutions orbital period radius vectors radius vector (Fig. 5.15) velocity energy per unit mass energy increment resulting from impulsive thrust [(Eq. 5.25)] turning angle (Fig. 5.7) flight angle (Fig. 5.16) latitude true anomaly, including integer multiples of 2n longitude gravitational parameter relative angular position angular velocity vector minimum periapsis propellant
Problems
()s ()sn
213
spacecraft synoptic
Problems (1) A spacecraft is on a 8000 km radius, circular orbit about Mars. A shortduration, impulsive thrust in the direction of motion is applied to increase the spacecraft's velocity further. Find numerically the m i n i m u m velocity increment that is needed to cause the spacecraft to escape from the Mars gravitational field. (The gravitational p a r a m e t e r of Mars is 42.81 103 km 3 / s2.) (2) To avoid an accumulation of space debris in the geostationary orbit, a spacecraft, having served out its usefulness, is being propelled into a new orbit. The orbit change consists of two short-duration, impulsive rocket motor burns. The thrust vectors are chosen to be either parallel or antiparallel to the direction of motion at the time. The first of the two thrusts, which is tangent to the geostationary orbit, is hence at the perigee (or apogee, if the thrust is retrograde) of the resulting elliptic orbit. The second thrust is at the apogee (perigee) of this orbit. The intent of the m a n e u v e r is to ensure that the final orbit is at a mini m u m distance of 300 km from the geostationary orbit. The final orbit can be either entirely on the outside or entirely on the inside of the geostationary orbit. The propellant c o n s u m p t i o n for removing the spacecraft will be proportional to the sum of the magnitudes of the two velocity changes A vx and Av2. Determine the magnitudes and directions of the two velocity changes for the case in which the sum IA vii + IA v2l of the magnitudes, hence the propellant consumption, is a m i n i m u m . (The earth's gravitational p a r a m e t e r is 3.9860 105 km 3/s2.) (3) Consider a H o h m a n n t r a n s f e r from a circular orbit of radius rl to a second, coplanar circular orbit of radius r2. Let A v the sum of the two velocity increments needed for the maneuver. Show that for a fixed gravitational p a r a m e t e r and radius rl the m a x i m u m required A v occurs for r2 -- 15.58rl. (4)* Consider a spacecraft that coasts on a m i n i m u m e n e r g y elliptic path in the gravitational field of the sun. The path starts at a point P1 in the vicinity of Jupiter. P1 is at 75.00 ~ ecliptic longitude, 1.00 ~ ecliptic latitude, and at a radius from the sun's center of 5.150 AU. It ends at point P2 in the vicinity of Pluto. P2 is at 335.00 ~ ecliptic longitude, -12.00 ~ ecliptic latitude and at a radius of 38.00 AU. (1 AU -- 1.495979 108 km; gravitational p a r a m e t e r of the sun = 1.32712 1011 kn33/$2.) (a) Compute the semi major axis and eccentricity of the path. (b) Compute the time required to travel from P1 to P2. (5)* A space vehicle is launched from the earth at sea level. The takeoffis vertical, followed by a maneuver, assumed instantaneous, that results in a flight angle of 88.0 ~ (i.e., 2.0 ~ from the vertical, the so-called "kick angle"). After this, the vehicle is to follow a gravity t u r n trajectory. The thrust and specific
214
C HA PT ER 5
Orbital Maneuvers
impulse, hence also the mass flow rate, are assumed constant throughout. Let ( )0 designate quantities at the time of initiation of the gravity turn. The following numerical data are assumed: Mass Thrust Specific impulse Flight angle Velocity
m0 = 400 000 kg Ft = 7.00 106N /sp = 350 s F0 = 88.0 deg. v0 = 300 m / s
(a) Write a computer program for the gravity turn trajectory in Cartesian coordinates. A reference frame fixed to the earth is assumed. The effect of the earth's rotation, earth surface curvature over the horizontal distance of the trajectory, and change of gravity with altitude are neglected. Continue the calculation to the point where the trajectory would be horizontal. (b) Modifythe program so as to include aerodynamic drag. The atmospheric density, pa, can be approximated by Pa = Psl
exp(-h/ho)
where h is the altitude, h0 = 9295 m the scale height of the atmosphere (near sea level), and psi = 1.226 k g / m 3 the standard atmospheric density at sea level. The m a x i m u m cross-sectional area of the vehicle is 18 m 2. The drag coefficient is assumed to be 1.0 throughout the gravity turn trajectory. Compare the result with that in (a). (c) Modify program (a) by formulating it in the more exact inertial frame represented by polar coordinates in an earth-centered, nonrotating reference frame. Launch in the equatorial plane is assumed. Include the change in gravity with altitude for a spherical earth. Compare the result with that in (a).
References 1.
Kaplan, M. H., "Modern Spacecraft Dynamics and Control," John Wiley & Sons, New York, 1976.
2.
Brown, C. D., "Spacecraft Mission Design," AIAA Education Series, American Institute of Aeronautics and Astronautics, Washington, DC, 1992. Prussing, J. E. and Conway, B. A., "Orbital Mechanics," Oxford University Press, New York, 1993. Chobotov, V. A., "Orbital Mechanics," 2nd ed., American Institute of Aeronautics and Astronautics, Washington, DC, 1996.
3. 4. 5.
Culler, G. J. and Fried, B. D., "Universal Gravity Turn Trajectories," Journal of Applied Physics, Vol. 28, No. 6, pp. 672-676, 1957.
6 Attitude Control The need for close control of the attitude (i.e., orientation in space relative to some frame to be defined in each case) of spacecraft follows from such requirements as the need to point antennas and sensors toward the earth or other astronomical objects. The required accuracy greatly depends on the particular application. It can be relatively low for broadcasting satellites. In some scientific applications the required accuracy of the attitude control may be 1 arc minute or less. Attitude control is also needed in the case of launch vehicles or upper stage vehicles, particularly because their orientation will influence the direction of thrust, hence the flight path. A useful distinction can be drawn between external and internal disturbances that trigger the attitude control. An example of an external disturbance is the effect that is produced on a spacecraft by solar radiation pressure. Usually the spacecraft configuration is such that it is asymmetric with respect to the sun line; consequently, there will be a torque. Such external torques can also be produced by gas leaking from a spacecraft, by aerodynamic torques in the upper atmosphere, and by the torque produced by the gravity gradient. In the case of a spacecraft that is not fully degaussed, the vehicle's p e r m a n e n t magnetic dipole, by interacting with a planetary magnetic field, will in general produce a small torque. In Fig. 6.1, rough estimates of the more important external torques acting on typical earth-orbiting spacecraft are shown. Above about 10,000 km altitude, the predominant external perturbation is usually caused by the solar radiation pressure. At a distance of 1 AU from the sun, this pressure is 4.4 10 - 6 N / m 2. For small spacecraft, or if the spacecraft has approximate symmetry, the solar radiation torque at this distance from the sun can be as small as 10 -S Nm or even less. External torques perturb the total angular m o m e n t u m of the vehicle. Ultimately, this must be corrected by a countertorque, such as can be provided by pairs of small thrusters. An example of an internal disturbance is the oscillating torque acting on the propellant tank, hence on the remainder of the vehicle, by fluid sloshing. Although this will not alter the vehicle's total angular m o m e n t u m , it will disturb the orientation of sensors, particularly attitude control sensors, that are m o u n t e d on the vehicle's shell. Particularly critical is also the attitude control of reentry vehicles and of vehicles that make use of aerobraking in a planetary atmosphere. Because the stream of propellant gas at the nozzle exit of rocket motors is highly turbulent, perturbation torques, m u c h larger than those shown in Fig. 6.1 for spacecraft, are produced each time such a motor is fired. On the 215
216
C H A P T E R 6 Attitude Control 10 5 _
,~. El04
10 3
l -
-
i
10 2 10 -6
Solar Radiation Torque Geostationary Altitude
i 10-5
Gravity Torque Magnetic torque Aerodynamic Torque I
10 .4 10-3 Torque (N m)
I
I
10-2
10-1
Figure 6.1 Orders of magnitude of the principal disturbance torques affecting earth-orbiting spacecraft. (Adapted from DeBra, D. B. and Cannon, R. H., "Momentum Vector Consideration in Wheeel-Iet Satellite Control," Guidance, Control and Navigation Conference, American Rocket Society, Stanford University, 1961.) other hand, the needed accuracy of control, typically of the order of 1 or 2 degrees, can be less in this case. In the case of launch vehicles, the main concern is usually the aerodynamic torque, which may be caused either by steering or by upper atmosphere wind shear. This effect can be critical because launch vehicles, particularly very large ones for which weight saving is at a premium, are not designed to tolerate substantial bending moments. In this chapter, reference will be m a d e mostly to spacecraft. Most results, however, translate easily to other vehicle types. Concerning attitude control, m a n y spacecraft can be classified into one of the two types represented in Fig. 6.2: spin-stabilized spacecraft or threeaxis stabilized ones. Spin-stabilized spacecraft (Fig. 6.2a) maintain their attitude by the gyroscopic effect resulting from spinning the cylindrically shaped portion of the spacecraft about its longitudinal axis. For full stability, the gyroscopic effect must be supplemented by so-called nutation dampers, passive devices that dissipate energy. The cylindrical portion is usually the more massive one because it typically contains such heavy items as the propellant needed for attitude control, the storage batteries, and often the empty case of the solid-propellant motor that was used for the final orbit insertion. Angular velocities of the order of one revolution per second are quite common. Spin-up is achieved by pairs of thrusters firing in a plane perpendicular to the nominal spin axis. The bore sights of the antennas and of some of the sensors need to be pointing in a desired direction, for instance, toward the earth. They are therefore m o u n t e d on a despun platform, which is supported by a shaft and bearings in the cylindrical section. The designation "despun" here refers to a reference frame corotating with the earth, not with inertial space. Since the angular velocity, relative to inertial space, of the platform is very low
6 Attitude Control A
B
9
.......
"
~
SolarPanels
Sun
iii ce''s
A/
(a)
217
Ea
t, u eCon ro,
Antennas---" (b)
"~~~
B
Figure 6.2 (a) Spin-stabilized spacecraft; (b) three-axis stabilized spacecraft. A-A, spin axis; B-B, solar panel rotation axis.
(for a circular orbit 2zr times the inverse of the orbital period), the despun section does not appreciably contribute to or subtract in this case from the gyroscopic effect of the spinning section. Electric power and signal circuits connect the two sections by means of slip rings. The bearing and slip ring friction is compensated for by small electric motors (as otherwise the two sections would start to share their angular momenta, speeding up the despun platform and slowing down the spinning section). For orbiting spacecraft, the nominal spin axis is perpendicular to the orbit plane. This allows the antennas on the despun section to keep on pointing in the desired direction as the spacecraft moves along its orbit. Except during maneuvers, spin-stabilized spacecraft do not require active sensors for their attitude control. The control system is therefore substantially simpler than that needed for three-axis stabilized spacecraft. On the other hand, because at any time only approximately one-half of the solar cells are illuminated, the m a x i m u m available solar electric power is more limited. Another reason for this is that the d r u m that carries the solar cells of spin-stabilized spacecraft is restricted in size by the launch vehicle payload space, whereas three-axis stabilized spacecraft can make use of fold-out solar panels, the size of which is not so limited. To carry out the maneuvers needed to reach final orbit, the entire vehicle is usually despun prior to its reorientation to match the intended thrust direction. This can be accomplished by firing a pair of small thrusters or by the yo-yo m e c h a n i s m described in Sect. 1.3. Having acquired the new orientation, the vehicle is spun up again. This serves to keep the attitude steady during the firing of the motor. Three-axis stabilized spacecraft (Fig. 6.2b) use actuators, such as thrusters, to maintain the desired orientation. For instance, in the case of
218
C H A P T E F~ 6 Attitude Control geostationary c o m m u n i c a t i o n satellites, the desired orientation is such that the communication antennas remain pointed toward the ground stations. Relative to inertial space, this requires one full rotation of the spacecraft once every 24 h. In the case ofdeep-space missions, the normal orientation ofthe vehicle is such that the high-gain a n t e n n a is directed toward the earth. An additional consideration is that the deployed solar panels should be close to being perpendicular to the sun line, so as to produce the maxi m u m solar electric power. For this purpose, they are m o u n t e d on a shaft, often on a c o m m o n shaft when there are two solar panels on opposite sides of the spacecraft. Relative to the spacecraft, this shaft is slowly rotated by an electric motor. For an orbiting spacecraft, the rate of rotation will correspond to the orbital period. With optimal orientation of the spacecraft, the m a x i m u m possible solar incidence angle on the panels will be 90 ~ at the time of spring and fall equinox and 66.5 ~ (90 ~ less the 23.5 ~ inclination of the ecliptic) at solstice. Figure 7.7 illustrates the astronomical and geometrical constraints. The twin requirements of pointing the a n t e n n a beams and at the same time the solar panels require that the orientation of all three spacecraft axes be controlled. This is accomplished by a control system that receives inputs from sensors and c o m m a n d s actuators. For orbiting spacecraft, the most important sensors are so-called horizon sensors that scan periodically the astronomical body that is being orbited. Several types of actuators are in use. C o m m o n ones are small thrusters arranged such that in various combinations they can rotate the spacecraft about any of its axes. Because they are peculiar to space technology, the emphasis in this chapter is placed on s e n s o r s a n d a c t u a t o r s . They are an essential part of all space vehicles. By contrast, the electronic aspects and the underlying c o n t r o l t h e o r y differ only in detail from their very broad applications in many other fields of technology. The goals, of course, are to achieve stability, quick response, and relative insensitivity to parameter changes. Control theory has become a discipline all by itself, usually taught in separate courses. It is therefore largely passed over in this book. The development of the software needed for attitude control is a major task. It is usually based on the description ofthe spacecraft rotations byEuler angles and their transformations. Because quaternions are mathematically the most natural way to describe rotations, it has also become customary to use them in software developments. This also goes beyond the scope of this book. The treatment in this chapter follows in major parts that of Kane, Likins, and Levinson [1] and that of Bryson [2]. Different treatments and additional material can be found in Refs. 3 to 6.
6.1
Principal Axes and Moments of Inertia of Spacecraft As a preliminary to attitude control analysis it is necessary to determine the principal axes and the principal m o m e n t s of inertia of the spacecraft, which, for this purpose, is assumed to be a rigid body. The importance of
6.1 PrincipalAxes and Moments of lnertia of Spacecraft
219
these concepts derives from their connection to the angular m o m e n t u m vector, say L, about the center of mass of the spacecraft. If r is the position vector extended from the center of mass, and w = w(t) the angular velocity of the spacecraft relative to inertial space, L = f r x (w x r ) d m
(6.1)
where the integration is extended over the total mass ofthe spacecraft. Making use of the vector identity r x (w x r)
= r2w
-
(w. r)r
a useful alternative expression for L is
t-
f r2 dm- f (., . r)rdm
We next introduce a set of mutually perpendicular axes with their origin at the center of mass and fixed to the spacecraft, hence corating with it. Let L'i, x[, o)i'(i - 1, 2, 3) denote the c o m p o n e n t s of L, r, w in this set of axes. Therefore I
I
f
I
!
I
I
f
f
I
f
f
f
f
I
L~ = I~1% + I12 % + I13 %
(6.2)
L~ = I~lO)l + 122092 + 123093 L~ = ~1(.01 -+- I32w2 + 1330)3 where
I lfl -
I:~2 =
f ( x 2 a + x 3t2) din,
,.:, =
1,,,-
f(X:32 -Jr-X[2) dm,
,=f
133
,2 /
(x~2 + x 2 ) d m
( i # j)
f x;x d m
(6.3)
More compactly, where ~ij is the Kronecker delta,
I~'j
f (*ijx/cx/c x[xj) dm
(6.3')
The terms I~1, I~2, etc. are the c o m p o n e n t s of the inertia tensor, say I. (That it is a Cartesian tensor, i.e., that it satisfies the transformation equations defining such tensors, requires a proof, not given here.) As is seen from (6.3), I is symmetric and has the matrix representation
Il
I13 I;z1
I
(6.4)
Equation (6.2) can be written more compactly as L = I~
(6.5)
Finding the principal axes a m o u n t s to determining the eigenvectors of the inertia tensor. By way of mathematical background, the reader will recall that the eigenvectors of a tensor T and the associated eigenvalues
220
C H A P T E R 6 Attitude Control are o b t a i n e d by considering the vector equation Tx = ~.x
(6.6)
If a vector x, other t h a n x = 0, satisfies this equation, the vector is called an eigenvector, say e, of T. Associated with each eigenvector e is its eigenvalue )~. Because the e q u a t i o n is h o m o g e n e o u s in e, there is no loss in generality by a s s u m i n g that all eigenvectors are of m a g n i t u d e one ("unit eigenvectors"). Equivalently, the e q u a t i o n can be expressed by a set of hom o g e n e o u s , scalar equations for the c o m p o n e n t s . For a nontrivial solution to exist, the d e t e r m i n a n t formed from the coefficients in these equations m u s t vanish, which gives rise to the so-called characteristic (or "secular") e q u a t i o n for the eigenvalues. Furthermore, if, as is the case for the inertia tensor, T is real a n d symmetric, it can be shown that all eigenvalues are real and that the eigenvectors can be chosen to be mutually orthogonal. Applying this to the inertia tensor, there will be three o r t h o n o r m a l eigenvectors e, each with its eigenvalue )~. The Cartesian c o m p o n e n t s of any chosen eigenvector in the coordinate system (x~, x~, x~) will be designated by el', e~, e~. The three scalar equations corresponding to (6.6) for this eigenvector are therefore (I~1-~.)e~
+ I1'2e ~ + I~3e ~
--
0
I~2e~ + (I22 -- k)e~ + I~3e ~
=
0
I~3e; + Ia3e2 ' ' + (I33 ' - Me 3'
-
0
(6.7)
The condition that there exist solutions other than the trivial one e = 0 therefore leads to the characteristic equation
I~l - ~. 1;2
1;2 I~2 - - ~.
I;3 I~3
113
I23
I33 -- )~
-- 0
(6.8)
hence a cubic in )v. As a c o n s e q u e n c e of the s y m m e t r y of the inertia tensor, all three solutions are real. They will be designated by ~ , )~2, )v3. Having d e t e r m i n e d the eigenvalues from the characteristic equation, the c o m p o n e n t s of each unit eigenvector in the (x'~, x[2, x~) coordinate syst e m are found successively from (6.7), together with the normalization condition I2
I2
t2
e1 + e 2 + e 3 - 1 If no eigenvalues are equal, we can choose a n e w Cartesian coordinate system (Xl, x2, x3), with origin at the center of mass, so that the unit eigenvectors just calculated are the base vectors of a n e w (right-handed) Cartesian coordinate system. The new axes are called p r i n c i p a l axes. As is easily shown, in this coordinate system the products of inertia vanish, so that the matrix representation of the inertia tensor in the principal axes system is
1/10 01 0
12
0
0
0
I3
(6.9)
6.1
PrincipaIAxes and Moments of lnertia of Spacecraft
221
where
11= f (x 2 +X 2) din,
12=
f (x 2 +x 2) din,
/3 =
f (x ff +x 2) dm q ]
o.I
tl
(6.10)
I1, 12, 13 are called the principal m o m e n t s of inertia. (A single subscript is used to distinguish t h e m from the m o m e n t s of inertia in the general case.) It follows easily that I1 = X~,
12 = ~.2,
/3 = X3
(6.11)
If two of the eigenvalues - - and therefore also two of the principal moments o f i n e r t i a - - are equal, the eigenvector associated with the third eigenvalue determines one of the principal axes. The other two principal axes are perpendicular to this axis and to each other but otherwise can be chosen arbitrarily. A circular cylinder with uniform density is an example. More generally, a geometrically irregular body, possibly with n o n u n i f o r m density, can still be an example if two of the eigenvalues are equal. One then speaks of an "inertially symmetric" body. To simplify the attitude control system, m a n y spacecraft are designed to be inertially symmetric. Often, this symmetry can be obtained solely by a judicious arrangement of the c o m p o n e n t s within the spacecraft envelope. Quite c o m m o n are c o m p o n e n t distributions that result in off-diagonal elements of the inertia matrix that are no more than about 1% of the principal m o m e n t s of inertia. If all three eigenvalues are equal, the orientation of the (mutually orthogonal) principal axes can be chosen arbitrarily, and all m o m e n t s of inertia are equal. The inertia tensor is said to be isotropic.
6.1.1
Application to Spacecraft Calculations In computing the principal axes and m o m e n t s of inertia of a spacecraft, the most time-consuming task is the calculation of the integrals in (6.3) for an a priori specified coordinate reference system (x~, x~, x~). For this purpose, the spacecraft is segmented into a large n u m b e r of parts whose contributions to the integrals are separately evaluated and then summed. Software exists that facilitates the segmentation and also has available the principal axes and principal m o m e n t s of inertia of m a n y geometrical shapes with constant or variable density, such as prismatic bodies, circular cones, or spherical shells. The spacecraft designer then need only specify the dimensions, the density, the location of the center of mass in the spacecraft reference coordinates, and the direction cosines of the part's reference axes with respect to the spacecraft reference axes. Two such segments, one of general shape and one prismatic, are illustrated in Fig. 6.3, together with their reference axes. Generally, the centers of mass of the segmented part and of the spacecraft will be different. So will the orientations of the respective reference axes. Therefore a transformation will be n e e d e d to make the segment's inertial tensor that is contained in the software useful for the calculation of the spacecraft inertial tensor. This transformation can be viewed as a two-step
222
C H A PT E R 6 Attitude Control "
X3n
,, x3~
x 1
(a)
(b)
Figure 6.3 Computation of the spacecraft inertia tensor from the inertia tensors of its components. Examples: (a) propellant tank; (b) electronic box. O, center of mass of spacecraft; On of nth component. process: (1) a parallel translation that brings the two centers of mass into coincidence and (2) a rotation that causes the two sets of reference axes to coincide. Let (x~, x~, x~) designate the spacecraft reference c o o r d i n a t e s of a point on the spacecraft, with their origin at the spacecraft center of mass. For convenience, one of the axes is usually chosen to be parallel to the nominal pointing direction of the principal sensor or of the high-gain antenna, or in the direction of the thrust axis. In general, the spacecraft reference axes will not be principal axes. (At the start of the design process, these are not known.) Further, let (xl," ~, x~i~, %,n "" ) designate the c o m p o n e n t reference coordinates of a point in the nth segmented part. The origin is chosen to coincide with its center of mass. Analogous to (6.3'), the (i, j)th c o m p o n e n t ]i'j,n of the inertia matrix of the part is ]ij,,, n -- f
r. [OijX k,,x k,, -- X i,, Xj,,1Jn d m
If rn is the radius vector from the of mass of the nth part, the (i, j)th after the parallel displacement of the coincides with the spacecraft center Ii'}, n = f
[3ij(rs + xs
(6.12)
spacecraft center of mass to the center c o m p o n e n t Ii'}, n of the inertia matrix, part's reference axes so that their origin of mass, is
+ x k" ) - (r i + x i" )(r~ + x j " )]n d m
The mixed terms that result from the multiplication vanish as a consequence of the definition of the center of mass, so that, with m n as the mass of the nth part Ii'}, n
=
]ij," n + m n
[(~ij rk' rk' -- ri ' rj'] n
(6.13)
Let l:.u,n = cos(X/'n', x~,~) be the direction cosines of the reference axes of the segmented part with those of the spacecraft. The (i, j ) t h element of the
6.2 The Euler Equations for Time-DependentMoments of lnertia
223
inertia matrix for the entire spacecraft, when expressed in the spacecraft reference system, is therefore lki, n~lj, 1' ntJkl, i r" _t_ mn[gkl rmr - 'm '
-jIi~i -- ~
rk' rlln )'
(i, j, k, l, m = 1,2,3)
n
(6.14)
This is the same quantity that was expressed earlier by (6.3), except now more detailed as needed for practical calculations. The computation required for finding the eigenvalues and eigenvectors of the spacecraft proceeds as in (6.8) and (6.7). As a simple example we m a y consider a prismatic body of constant density. This is often a useful approximation for electronic b o x e s on spacecraft (Fig. 6.3b). In this example it is convenient to define the reference axes of the part to coincide with its principal axes. If the length, width, and height are bl, b2, and ba, the corresponding m o m e n t s of inertia are easily found to be
]ij" = tl O(mb/12)(b2 + b2 + b2 - b2)
if/=if i r jj
where mb is the mass of the box. Hence from (6.13)
,, Iij =
/ [
mb b~ + b 2 + b~ 12
]
b 2 + r~2 + r;2 + r~2 _ r i,2
- mb r~r)
if i = j if i r j
The contribution by the box to the spacecraft inertia tensor, expressed in the spacecraft reference system, then follows from (6.14).
6.2
The Euler Equations for Time-Dependent Moments of Inertia In analyzing the attitude control of spacecraft, it can usually be assumed that the inertial properties are time independent. Although the properties will change as propellants are expended, this change is usually very slow compared with the attitude control frequencies and can be neglected. The attitude control of such spacecraft can then be described by means of the classical Euler's equations for the rotational motion of rigid bodies. In some other, more rare, instances the temporal change of the moments of inertia needs to be taken into account. This can occur with small, rapidly spinning vehicles with fast-burning solid propellants at the time of the main motor firing. As propellant is expended, the center of mass moves relative to the rest of the vehicle, a n d - more i m p o r t a n t - the m o m e n t s of inertia change. As will be assumed in the following, in these applications the orientation of the principal axes relative to the vehicle will not change. In what follows, we derive the modifications of Euler's equations that are needed when the time rates of change of the m o m e n t s of inertia are comparable to the rate of rotation of spinning vehicles. In practice, the case of interest is the inertially symmetric vehicle with nominal spin axis, thrust
224
C HA PT ER
6
Attitude Control
axis, and one of its principal axes all coinciding. This axis will be designated byx3. The angular m o m e n t u m equation in the form most useful for applications to spacecraft refers to the total angular m o m e n t u m L of the spacecraft about its (instantaneous) center of mass. If M is the resultant m o m e n t (such as may result from the disturbance torques shown in Fig. 6.1) about the center of mass of the spacecraft and caused by external forces acting on it, d L / d t -- M
(6.15)
Here, and in what follows in this chapter, d()/dtrefers to the time derivatives in inertial space, whereas (') refers to time derivatives in the space fixed to the spacecraft. (In addition to external moments, there can also be internal moments, such as those caused by the sloshing of liquid propellants, that disturb the attitude; these do not affect the total angular momentum.) From (1.1) for the connection between the two time derivatives, dL/dt=
L +~ x L
where ~ is the instantaneous angular velocity of the spacecraft relative to inertial space. In c o m p o n e n t form, expressed in the spacecraft principal axes coordinate system (x~, x2, x3), L1 = Il a~l,
L2 = I2a~2,
L 3 = 13033
as follows from (6.2) when applied to the principal axes. From this follows the final result 11&l + Wl il + (13 - I2)a)2093 -- M1 / I2dJ2 d-~02J2 -~- (11 - 13)033091
--
M2
I 3 ~ 3 -t-r
--
M3
-~- (/2 -- I1)WlW2
J
(6.16)
It may be noted that the vector w is special in the sense that there is no distinction between its derivative in the two reference spaces. For it follows immediately from (1.1) that d w / d t = &. Strictly speaking, the components Wl, w2, w3 are not simply time derivatives of certain angles. Instead, they are nonintegrable combinations of time derivatives of the angular displacements. The second term in (6.16) represents the effect caused by the rate of change of the principal m o m e n t s of inertia. When these rates can be neglected, the well-known Euler equations II&l + (13 -- I2)a~2a~3 -- M1 /
/ M3
I2d~2 -~ (I1 -- I3)a~3~01 = M2 I3dJ3 +
(/2-
11)(-o1(-o2 --
(6.16')
for the rotational motion of solid bodies result. When the rates of change of the inertial terms are small compared with the attitude control frequencies, as happens in the majority of attitude control problems, the inertial terms can be treated as quasi-constant. That is, Eq. (6.16') can be applied at each instant, although the constants change on the (longer) time scale that characterizes the rate of propellant depletion.
6.3 The Torque-Free Spinning Body
225
In applications, most often the m o m e n t s of inertia and the external torques are given, and the task is to find the angular velocity. Although there are important special cases in which the equations are linear and simple to solve, in the general case Euler's equations are nonlinear.
6.3 The Torque-Free Spinning Body As an example of the application of the Euler equations (6.16') we consider the rotational motion of a rigid, spinning body. This topic is treated in most textbooks in mechanics. Because it has n u m e r o u s applications to spacecraft, it is also included here.
6.3.1
The Torque-Free, Inertially Symmetric, Spinning Body Spin-stabilized spacecraft are usually designed so that two of the principal m o m e n t s of inertia, say 11 and 12, are very close to being equal. Such spacecraft are said to be inertially symmetric. In practice, the agreement m a y b e 1% or even better. This can be achieved byjudiciously placing within the envelope of the spacecraft its various components. Because most spacecraft are more Weight than volume limited, the designer has considerable freedom in placing the components to achieve this symmetry. The attitude control of inertially symmetric, spinning spacecraft is substantially simpler than for asymmetric vehicles. Let Xl, x2, x3 be the body's principal axes and Wl, w2, w3 the corresponding of the angular velocity components relative to inertial space. Rigid bodies for which one of the principal m o m e n t s of inertia, say/3, is smaller than the two (equal) others are called prolate (Latin latus: side, flank; literally "protruding flank," i.e., pencil-like). If larger, the body is called oblate (literally "shortened flank," i.e., disklike). The axis x3, with m o m e n t of inertia /3, is usually the nominal spin axis. The two cases are illustrated in Fig. 6.4. Assuming that I1 = /2 and that there are no external torques, Euler's rigid body equations simplify to
d91+(~~31 - 1 ) 092093 - 0 &2-
(~1I 3- 1)
WlW3 = 0 &3 -
(617)
0
As indicated by the last of these equations, oJ3 is constant. The first two equations therefore are a pair of linear differential equations for 0~1 and oJ2. Let a =
~1 - 1
0)3
(6.18)
therefore also a constant. With this definition, the general solution of (6.17) is seen to be 0)1 - -
0)12 COS(~'~t --
X)
m2 -
w12 s i n ( f 2 t -
X)
[
/
(6.19)
226
C H A P T E R 6 Attitude Control
Figure 6.4 Torque-flee, inertially symmetric, spinning body. (a) Prolate body (11 = 12 > 13, unstable motion); (b) oblate body (11 = 12 < 13, stable motion). O, center ofmass; L, angular momentum; ~o, angular velocity; xl, Xz, x3, body-fixed, principal axes; O, nutation angle. w h e r e x is an arbitrary p h a s e angle a n d 0~2 the a m p l i t u d e . It follows that oJ2 + oJ2 = ~o22 - const. which shows that ~o~2 is the projection of ~o on the (x~, x2) plane. The projected vector, ,~2, is s e e n to rotate in time a r o u n d the x3 axis at the rate ~2. Also, since ~0~2 a n d ~o3 are b o t h constant, the m a g n i t u d e , ~o, of the angular velocity r e m a i n s constant. For the s a m e reason, the angle, say a, b e t w e e n the angular velocity vector a n d the x3 axis stays constant. Because, by a s s u m p t i o n , there is no external torque, the angular m o m e n t u m , L, is c o n s t a n t relative to inertial space (although not relative to a body-fixed reference). F r o m (6.2), w h e n specialized to principal axes, a n d with/1 = 12, it follows t h a t L1 = 11o31,
L2 = Il oJ2,
L3
-
-
"
]'30)3
The projection L]2 on the (x~, x2) p l a n e of L is therefore collinear with the projection of w. The a n g u l a r m o m e n t u m , the a n g u l a r velocity, and the x3 principal axis of the b o d y are therefore all in the s a m e plane. Given that initially the x3 axis includes an angle 0 with L, the x3 axis will precess a b o u t L at a c o n s t a n t rate a n d c o n s t a n t 0. This angle is often called the n u t a t i o n a n g l e ( n u t a t i o n = n o d d i n g motion; the expression stems from the m o r e general case in which the three principal m o m e n t s all differ a n d 0, c o n t r a r y to the inertially s y m m e t r i c case, can be oscillating, h e n c e the "nodding").
6.3 The Torque-Free Spinning Body
227
There exists a simple relation between the two angles a and 0. Since, as is readily seen, tan ~ = ~o12/~o3and t a n 0 = L 12/L3, it follows that t a n O = (11118) tanc~
(6.20)
Hence in the prolate case a < O, in the oblate case c~ > 0. As illustrated in the figure, the angular velocity vector can be visualized as being coincident with the contact line between two circular cones, one centered on the x3 axis and fixed to the body. It is called the b o d y cone. The other cone is centered on L, is fixed in inertial space, and is called the space cone. It is easily shown that the two cones roll on each other without sliding. In the prolate case, the two cones are exterior to each other; in the oblate case, the body cone is exterior to the space cone. Different initial conditions merely change the half-angles of the two cones. The same system of axes (Xl, x2, x3) is also shown in the lower part of the figure, together with the Euler angles and their time derivatives. The angular velocity vector, in place of decomposing it into its components ~Ol, co2, ~o3, can also be represented as the vector s u m of a vector p parallel to the angular m o m e n t u m vector L and a vector s parallel to the body symmetry axis, x3. The first is referred to as the precession, the second as the spin. The magnitude of p is the rate of precession and measures the rate at which the plane that contains L, w, and the x3 axis rotates about L. Given the m o m e n t s of inertia, there are two i n d e p e n d e n t variables, for instance, ~o3 and 0. All other quantities of interest can be expressed in terms of these. For instance, from p sin 0 = ~o3 tan c~, together with (6.20), follows
13 co3
(6.21)
P - I1 cos 0
6.3.2
Stability of the Motion Even in the absence of external torques, spinning spacecraft can slowly lose kinetic energy of rotation. Propellant sloshing and friction in the bearings between the spinning part and the d e s p u n platform can convert some of t h e kinetic energy into thermal one. The total angular m o m e n t u m will be unaffected, but the nutation angle can slowly change. This consideration leads naturally to t h e question of stability of the gyroscopic motion. In an analysis that is at least qualitatively correct, one assumes that at each instant the motion can be described as in Sect. 6.3.1 and that there is merely a slow change of the parameters. (The notion of folding all dissipative effects into a single scalar quantity that acts to diminish the rotational kinetic energy of the spacecraft is being used frequently. It appears to be impossible to establish this entirely from first principles. Indeed, some conditions, although pathological in practice, have b e e n discovered where this breaks down. In realistic problems of spacecraft motions no such anomalies have ever been found.) The kinetic energy of rotation is T = ~ 1 f (w x r) 2 d m
(6.22)
228
C H A P T E R 6 Attitude Control where r is the position vector extended from the center of mass. Evaluating the vector p r o d u c t and using the definitions (6.3) of the m o m e n t s and products of inertia result in 1 ! -~o) i! O)j! Iij,
T-
i, j -- 1, 2, 3
Writing (6.2) for the angular m o m e n t u m more compactly, !
L ; - Ii~o J,
i, j -
T-
~w.1 L
1,2,3
Hence (6.23)
As is readily seen from Fig. 6.4, the projection ofthe angular m o m e n t u m on the x3 principal axis is L 3 -- L cos 0. Also L 3 = I30~3 -- I3oJ cos c~. Therefore L cos 0 09 --
h cos
Writing in (6.23) the scalar p r o d u c t in terms of the principal axes components and using the expression just found for oJ, T =
L2
cos0 cos(0 - a)
9
213
cosc~
(6.24)
The factor containing L a n d / 3 is constant. The second factor containing the cosines, w h e n e x p a n d e d in a power series about 0 = 0, results in T-
213 1 -
1-~
02
(6.25)
valid approximately for small n u t a t i o n angles. The factor 1 - 13/11 is positive for prolate, negative for oblate bodies. As the rotational kinetic energy decreases because of internal dissipative effects, 0 is seen to increase in the prolate, decrease in the oblate case. A p r o l a t e body, if initially spun approximately about its inertial axis o f s y m m e try, will tend to increase its n u t a t i o n angle, hence is unstable. It will finally
x3
Initial
x3
~x
3
Final
Figure
6.5
Unstable motion of spinning body with h =/2 > 13.
6.3 The Torque-FreeSpinning Body
229
go into a flat spin, as illustrated in Fig. 6.5. On the other hand, an oblate body will tend to return to a spin about its inertial axis of symmetry, hence be stable. The rate of dissipation of kinetic energy in spacecraft is low. The growth of the instability of prolate vehicles is therefore correspondingly slow. In practice, it may take m a n y minutes before it reaches amplitudes that could be of concern. Nutation dampers, discussed in Sect. 6.6.2, are used to eliminate entirely the instability of spinning, prolate vehicles. For aerodynamic and structural reasons, the diameter of the shroud that contains the spacecraft on the launch vehicle is typically smaller than its length. Therefore this often calls for prolate spacecraft. Nutation dampers therefore find frequent applications.
6.3.3
General Case of the Torque-Free, Spinning Body The analysis of the general case of a torque-flee, spinning body with principal m o m e n t s of inertia that all differ from each other is considerably more complicated than the inertially symmetric case. One reason is that the simple reduction of the solid-body Euler equations to a linear set is no longer possible. This general case forms a classical topic in mechanics. It was studied by Poinsot (1777-1859), Klein (1849-1925), and others. The solution can be expressed in terms of elliptic integrals. But with the advent of high-speed computers it is m u c h more practical to integrate Euler's equations directly for the parameter values of interest to the spacecraft designer. The motion, relative to inertial space, of such a body can be viewed as consisting of three parts: (1) a rotation about one of the principal axes; (2) a precession of this axis about the (constant) angular m o m e n t u m vector; (3) a nutation (nodding) of the axis relative to the angular m o m e n t u m vector. The motion turns out to be stable if the body is initially spun about the principal axis that corresponds to the m a x i m u m m o m e n t of inertia (or, more generally, if the angle between the angular velocity and the axis with the m a x i m u m m o m e n t of inertia is less than 90~ Although spin stabilization of inertially nonsymmetric spacecraft is feasible, few such vehicle designs have been considered. When physical and geometric requirements imposed by the payload dictate a nonsymmetric design, three-axis stabilized (nonspinning) spacecraft are usually preferred. It has h a p p e n e d on occasion that because of a faulty deployment of a spacecraft component, or because of an error in the attitude control system, a spacecraft was sent into an uncontrolled t u m b l i n g motion. Often, the control can be recovered and an otherwise catastrophic event prevented by first deducing from telemetry the character and magnitude of the tumbling motion, followed by computer modeling of the event, using the full Euler equations. This allows one to examine the consequences of proposed corrective actions. The final step is then to initiate the selected remedial actions by firing the correct thrusters in the correct sequence and for the correct duration.
230
C H A P T E R 6 Attitude Control
6.4
Attitude Control Sensors To control the attitude of spacecraft, reliance is placed on sensors ofvarious types. This is particularly the case for three-axis stabilized vehicles but also applies to spin-stabilized vehicles. Sensors that are directed toward a star or toward the earth's horizon depend on sources external to the spacecraft. Others, such as gyroscopes, are internal. In all cases, an error signal is generated that then is used by the attitude control system to initiate the needed correction, for instance, by firing the appropriate attitude control thrusters. Usually, the attitude of a spacecraft is defined relative to inertial space, that is, relative to axes that do not rotate. For orbiting spacecraft, it is sometimes preferred to relate the vehicle attitude to the type of reference frame familiar from aircraft attitude control by introducing the pitch (relative to the local horizontal), roll, and y a w angles. It should be noted, however, that as the spacecraft moves on its orbit, the local horizontal rotates. Hence allowance must be made for the fact that the aircraft-type reference is not an inertial reference.
6.4.1
Star Sensors
In deep-space missions, as opposed to earth or planet orbiting missions, the spacecraft attitude is usually determined by observations of stars by small, optical telescopes. The attitude is then specified by the orientation of a set of spacecraft-fixed reference axes relative to the three mutually orthogonal axes that can be defined by the ecliptic plane and the vernal equinox line (Fig. 1.1). For this purpose, the declination and right ascension (Fig. 1.4) of selected guide stars are programmed into a star catalog contained in the memory of the flight computer. To fix the attitude, a m i n i m u m of two bright stars ("guide stars") need to be sighted. For high accuracy, the angular separation between the two stars should be not m u c h less than 45 ~. To avoid locking on the wrong star, the immediate vicinity of the selected guide stars should be void of other bright stars. A frequently selected star is Capella. It satisfies these conditions and also has the advantage of a high declination. For deep-space missions close to the ecliptic plane, as most such missions are, a high declination avoids optical interference from the sun or from the planets. In addition to the declination and right ascension of the guide stars, the flight computer should contain in its m e m o r y the positions of other bright stars. This greatly facilitates the reacquisition of the guide stars in case lock-on has temporarily been lost. The attitude of the spacecraft is therefore determined by the lines of sight to two guide stars relative to spacecraft-fixed reference axes. The geometric relation between the selected astronomical reference (e.g., based on the ecliptic plane together with the vernal equinox line) and the spacecraft reference axes can then be specified either by the direction cosines between the two systems of axes or by their Euler angles.
231
6.4 Attitude Control Sensors
Alternatively, a single guide star together with the sun may be used. In deep-space missions, the direction in the heliocentric reference of the spacecraft-to-sun line changes depending on the position of the spacecraft along its trajectory. Hence, in this case the position of the spacecraft at any given time has to be known.
6.4.2
V-Slot Sun Sensors Sun sensors are used to find the declination, say S, of the sensor axis relative to the heliocentric reference. The right ascension cannot be d e t e r m i n e d from a sun sensor. Therefore, to determine the spacecraft attitude completely, additional information, for instance from a star sensor or from a gyrocompass, is needed. Most sun sensors are made to rotate continuously by an electric motor. Alternatively, an oscillating mirror can be used to scan the sun. In the case of spin-stabilized vehicles, the needed rotation can be provided by attaching the sensor directly to the spinning part. The principle of operation is explained in Fig. 6.6. Figure 6.6a is a view parallel to the sensor axis of rotation. Figure 6.6b is the projection on the plane containing the sensor axis of rotation and the perpendicular to the ecliptic plane. The functioning of the sensor can best be explained by basing it on the geometry of a sphere with center C on the axis of rotation. For this reason, the sensor is depicted in the figure as a sphere of unit radius. In practice, the sensor outer surface could just as well be a fiat surface. There are two slots, $1 and $2. When the sun is in the plane C-S1, a portion of the light will strike the photocell P~, which is located on the line C-A1 approximately through the midpoint of $1. Hence the photocell is (a)
A1
S1
A2
0
~I.D
Irace of Sun7 / /
t2 .
-
/:/
::/;~
~
~"
--
ptic Pla ~A 1 =cos8
0
(c)
\
Figure 6.6 Sun sensor represented by unit sphere. (a) Top view; (b) elevation; (c) spherical triangle used for calculation.
232
C H A P T E R 6 Attitude Control
pulsed once on every revolution of the sensor. The slot 82 and the photocell P2 function analogously. The two slots are arranged in the form of a letter V. Slot $1, but not $2, is in a plane that contains the axis of rotation. As the sensor rotates, the path of the sun as seen by the unit sphere is incident on it on a plane parallel to the sphere's equator. The declination, ~, of the sensor axis of rotation in the heliocentric reference is the angle in the diagram between the axis of rotation and the intersection of the plane of incidence with the unit sphere. The period, say P, of the sun sensor, relative to inertial space, is equal to the time interval between two successive pulses ofphotocell P1. By averaging over m a n y rotations, P is determined to high precision. If t~ is an instant of time w h e n P~ is triggered, and t2 the instant following tl w h e n P2 is triggered, then the azimuthal angle, q~, traversed by the sensor during the interval t2 - tlis -- (2rc/P)(t2-
tl)
(6.26)
The angle ~ will be needed to find ~. This is seen from considering the spherical triangle shown in Fig. 6.6c. The angles a and fl are geometrical properties of the sensor, hence are known. Also known, from (6.26), is ~. The remaining side, a, of the spherical triangle can be eliminated as shown next: From the two Napier formulas of spherical trigonometry,
tan
+ a cos(~ - fl)/2 a + zr/2 -tan 2 cos(~ + fl)/2 2
and tan ~
-a 2
=-
sin(~ - fl)/2 c~ + 7r/2 tan sin(q~ + fl)/2 2
The desired result for the declination 8 of the sensor axis follows from applying the inverse tangent function to both sides of each equation and adding, with the result that sin(4~ + fl)/2 cotanc~
- tan -~ cos(4~ cos(4~ + - fl)/2 fl)/2 cotanc~
16.27)
The declination can therefore be found from the measured time intervals between photocell pulses and the application of software that is prog r a m m e d to express Eqs. (6.26) and (6.27).
6.4.3
Horizon Scanning Sensors Orbiting spacecraft often use as an input to their attitude control system observations of the h o r i z o n of the orbited astronomical body. The required sensors usually operate in a scanning m o d e in which the sensors' lines of sight are swept repeatedly over the astronomical body such that they cross its horizon. In the case of three-axis stabilized spacecraft, the scanning motion is achieved either by a continuous conical sweep of the sensors' lines of sight
6.4 Attitude Control Sensors
233
Figure 6.7 Horizon scanning sensor (only a single line-of-sight cone is shown), ta, start; fa, end of intersection with earth; 0, pitch angle; r roll angle.
or by motor-driven mirrors that oscillate back and forth through some finite angle. In spin-stabilized spacecraft the sensors can be rigidly attached to the spinning portion of the vehicle. In this case the vehicle itself provides the sweeping motion of the lines of sight. The basic principle of operation of such sensors is illustrated in Fig. 6.7. In this figure the earth is taken as an example. The line of sight moves on a circular cone that intersects the earth during part of the time. As is illustrated in the figure, the intersection starts and terminates on the horizon as seen from the spacecraft. On a given sweep the intersection starts at some time tA and terminates at t~(. It is the difference between these two times that provides the information needed for the attitude control. In the usual arrangement, illustrated in Fig. 6.8, there are two lines of sight, A and B, provided either by a single sensor or by two separate but synchronized sensors. The vehicle's pitch angle (relative to the local horizontal) is designated, in conformity with the usual aircraft terminology, by 0, the roll angle by r and the yaw angle (not shown) by ~. It is evident that because of the spherical symmetry of the earth, ~ cannot be determined by horizon sensors. To obtain it requires information derived from gyroscopes. (More precisely, the outputs of two rate gyros, one for the roll rate and one for the yaw rate, are c o m b i n e d with the Kalman filtered roll angle obtained from the horizon sensors.) The photodetectors and optics of horizon sensors are designed to operate in the near-infrared, usually in the wavelength range 10 to 20/zm. The advantage is that in this spectral range the earth's atmosphere is more sharply defined. Also, there is less variation of the signal's amplitude between a sunlit and a dark spacecraft horizon.
234
C H A P T E R 6 Attitude Control
'
Local Horizontal
o,, x,s
/t;
tB
tA
Earth Center
Figure 6 . 8 Horizon scanning sensor with dual lines of sight A and B. Dashed lines; earth trace at zero spacecraft roll; solid lines; general case. Let rg be the earth's radius (averaged for the orbit and including a correction for the effective height of the atmosphere in the near-infrared region) and r the distance of the spacecraft from the earth's center. The radius, rH, of the horizon seen from the spacecraft is therefore rH -- r g v / 1 - q / r 2
(6.28)
Figure 6.8 shows the ground traces of the two lines of sight A and B for a given (here positive) roll angle ~. Shown by dashed lines for reference are the ground traces for zero roll angle. These are symmetric with respect to the plane defined by the roll axis and the local vertical. The angle, say 2/~, between the two sensor telescopes with the lines of sight A and B in principle can be chosen arbitrarily as long as A and B sometimes intersect the earth. For comparable sensitivity of the pitch and roll angle determinations it is, however, advantageous to choose ~ such that for zero roll angle the four points oftangencywith the earth form approximately a square. This is also shown in the figure. It then follows that the preferred choice for ~ is /~ - tan -1 (1 + 2r 2 4 / r g ) -1/2
(6.29)
Let y (t) be the (positive) angle between the spacecraft roll axis (i.e., the reference axis for the pitch) and the plane containing the two lines of sight. This angle can be determined by a shaft encoder on the axis of rotation of the sensor.
6.4 Attitude Control Sensors
235
As the plane is swept over the earth, there will be an instant, say tv, w h e n this plane is vertical. This will occur at the time 1
1
t
tv = ~(tA + t~) -- ~(tB + tB)
(6.30)
The value of ~, at this time is sufficient to d e t e r m i n e the pitch angle 0 of the spacecraft, since evidently
(6.31)
0 = y(t = tv) - 90 ~
The period of the sweep is of the order of a few seconds, which is short c o m p a r e d with the time for appreciable changes of the spacecraft's pitch. The m e a s u r e m e n t s of F at the midpoints b e t w e e n ta and t~( and similarly b e t w e e n tB and t~, averaged for greater precision as indicated in (6.30), in effect provide a quasi-continuous reading of the pitch angle. The roll angle can be d e t e r m i n e d from the time intervals t~( - tA a n d t~ - tB. This is seen from Fig. 6.8 as follows: For roll angles r ~ 0, the lines of sight, w h e n just touching the horizon, b e c o m e displaced by c o m p a r i s o n with the case r = 0, resulting in a lengthening or shortening of the ground traces, h e n c e of the time intervals. If (At)0 is the value of the time intervals for zero roll angle, it follows from the geometrical relations shown in the figure that
r
2(1+2r2/r~) t~ -
tA - -
{At)0
t~ -
tB -
(At)0-
+
.
-
-
1+
2(1 +
, .
1 +2r2r~t/r~
-
v
valid for [r << 1. The angular rate of the sensor axis of rotation (or of the rate of spin in the case of spin-stabilized vehicles with sensors fixed to the vehicle) is designated by v. Subtracting the two equations and solving for r gives the final result r -
(1 + 2r2r2/r~)v (t/~- t A - (t;3 - tB)) 4(1+2r2/r~)
(6.32)
Measuring the time intervals t~( - tA and tA - tB a n d carrying out the calculations indicated in (6.28) and (6.32) therefore d e t e r m i n e s the roll angle. This supposes that the orbit radius, or m o r e generally the i n s t a n t a n e o u s spacecraft-to-earth-center distance, r, is known, derived either from k n o w n orbital data or by a spacecraft position determination. A particular design of a horizon scanning sensor is shown in Fig. 6.9. The (single) line of sight is through an infrared-transmitting g e r m a n i u m window, coated with an interference filter to define the passband. A wedges h a p e d lens, rotated by an electric motor, directs the incident radiation toward a bolometer. In this way a conical scan of the earth is obtained. The angle of rotation, a n d from it the angular rate, is obtained from magnetic pickups. Infrared radiation from spacecraft c o m p o n e n t s that m a y intrude into the scanning cone are electronically blanked out. Two such instruments, working in t a n d e m , are n e e d e d to complete the system illustrated in Fig. 6.8. Pitch a n d roll angle accuracies achieved are of the order of 0.1 ~.
236
C H A P T E R 6 Attitude Control
Figure 6.9 Schematic of a horizon scanning sensor. Courtesy of TELDIX GmbH, Germany. 6.4.4
Rate and Integrating Gyroscopes Most spacecraft are equipped with gyroscopes of several different types. In particular, gyroscopes can be used for the determination of the orientation in inertial space of the chosen spacecraft reference axes and for the corresponding rates of change. Modern gyroscopes, either of the mechanical type or laser-ring gyros, are the product of a long development and have reached unprecedented precision. A large body of technical literature exists, m u c h of it applicable to space technology. The present discussion omits all design aspects. Figure 6.10a illustrates schematically a two-gimbal gyro. The rotor, which is driven at a high and precisely controlled speed by an electric motor, is supported by bearings in the inner gimbal. This gimbal is supported by the outer gimbal, which in turn is supported by the reference frame. In some applications, the reference frame is rigidly attached to the spacecraft ("strap-down" gyro); in others it is supported by a platform, which is attached to the spacecraft by gimbals. Sometimes the platform orientation is controlled by feedback control from the gyros so as to be nonrotating relative to inertial space. Sometimes, in orbiting spacecraft, it is preferred to control the platform to remain at all times locally horizontal. Gyros used for attitude control are m a d e to be insensitive to translational accelerations by having the centers of mass of rotor and gimbals coincide with the geometric center, that is, with the c o m m o n intersection of rotor and gimbal axes. Some gyroscopes, referred to as rate gyros, measure one or several components of the spacecraft's angular velocity. Others, referred to as integrating gyros, are designed to measure the total angle through which the spacecraft has rotated about a specified axis in a given time interval. Integrating gyros therefore provide time integrals of angular velocities. In what follows, the discussion will be limited to single-gimbal gyros. Gyros ofthis type are schematically represented in Fig. 6.10b. The (common) principal axes of rotor and gimbal are designated by x~, X2, X3. The
237
6.4 Attitude Control Sensors x2" ~ - - InputAxis -- OuterGimbal --Inner Gimbal B
I % Rotor~
['~ "~"~
OutputAxis
B J
,
X 1"
Gimbal
~
ReferenceFrame~
x3
(a)
(b)
F i g u r e 6.10 Schematics of (a) two-gimbal gyro and (b) single-gimbal rate gyro. xl, x2, x3, principal axes of rotor and inner gimbal; v, rotor rate of rotation; N1, restoring and damping torque on gimbal; S2, spacecraft angular velocity component; B, bearings.
axes x;, x~, x~, also mutually orthogonal and through the c o m m o n center of mass, refer to the reference frame and are parallel to spacecraft reference axes (not necessarily spacecraft principal axes). Axes Xl and x~ coincide. The corresponding angular velocity c o m p o n e n t s relative to inertial space are 091 (t), 092(t), v -k- 093 (t) for the rotor; O91 (t), 092(t), 093 (t) for the gimbal; and oJ~ (t), oY2 (t), co'3 (t) for the reference frame. Here v is the angular rate of the rotor relative to the gimbal. When the spacecraft rotates, the rotor will force the gimbal to rotate about its Xl axis, resulting in a change of the angle 0 between gimbal and reference frame. In rate gyros a proportional restoring torque - k O ( k = spring constant) and d a m p i n g torque - c O (c = d a m p i n g constant) are applied from the reference frame to the gimbal. This torque can be visualized as being supplied by a torsion spring and mechanical damper. More precisely than by mechanical means, the torque is obtained electronically. In integrating gyros, k = 0. As is evident from the figure, the shafts and bearings impose kinematic constraints. They relate w and w' by !
091 - - 091 + b ,
092 = o92' c o s 7) -Jr- 093'
sin0,
093 = --09 2' s i n 0 + 093' COS L9
(6.33) The Xl, x2, x3 c o m p o n e n t s of the angular m o m e n t u m , L, c o m b i n e d for rotor and gimbal are L1 = (h + 11)~1, L2 = (h + J2)w2, L3 -- /3(1; q- 093) q- ]3093
(6.34)
where h = 12 and 13 are the transverse and axial m o m e n t s of inertia of the rotor and 11, ]2, 13 the m o m e n t s of inertia of the gimbal. Designating time derivatives in inertial space by d( )dt, and in the space defined by the gimbal
238
C H A P T E R 6 Attitude Control
by ('), from (1 1) in Chap. 1, d L 1 / d t = L 1 + a)2L3 - (.03L2 -- N1
(6.35)
where L I - (I1 + ]1)(-01 and the torque N1 = - k O - cO. From (6.33) and CO1 -" COt1 -Jr- ~ -- d~o ] ~dr + ;)
[again an application of (1.1)], this becomes (I1 + I1)
-~
+ 0
+ 13 (v + ~03) o)2 + ]3o)2o)3 - (I2 + 12)0)2003 - - k O - c3
(6.36) In the applications of interest to attitude control, the gyro rotor rates of rotation are very high in comparison with spacecraft angular velocities. The rotor may spin at 50,000 r p m or higher, whereas even a spin-stabilized spacecraft will hardly exceed 50 rpm and a three-axis stabilized vehicle m u c h less. Similarly, the relative time rates of change of spacecraft rates ofrotation will be m u c h less than the rotor rate ofrotation. Expressed in terms ofstrong inequalities, Iw'l << v
and
I d w ' / d t l << vlw'l
(6.37)
The terms that contain wzw3 or d(co])/dt as factors are therefore negligible in comparison with the term containing v, so that to a very high degree of approximation (I1 + J 1 ) 0 + cO + kt9 = - I 3 Y w 2
(6.38)
The h o m o g e n e o u s part of this equation is seen to describe a harmonic, d a m p e d oscillation with the natural frequency COn = v/k/(I1 + 11)
and the d a m p e d frequency w~ - ~OnV/1- (2 where ( is the nondimensional damping coefficient defined by ( = C/2COn(I1 + 11)
Up to this point, the development applies to both rate and integrating gyros. Considering now the case of rate gyros, k and c are m a d e large [with ( = O(1)] so that the time for damping the gimbal becomes m u c h shorter than the time for significant changes of the spacecraft attitude rate. Therefore, on the slow time scale of co2, the pseudostationary solution O(t) = -
I3v k
~o2(t)
(6.39)
applies. Another consequence of a large spring constant is that the gimbal deflection is small, limited only by the need for a precise electronic readout of O. The difference between o)2 and co~ can then be neglected. If ~2 is the c o m p o n e n t of the spacecraft angular velocity in the direction of the x~ axis,
6.4 Attitude Control Sensors
239
referred to as the gyro's input axis, the main result characterizing rate gyros becomes ~2(t) =
kO(t)
I3v
(6.40)
The result is seen to be i n d e p e n d e n t of the gimbal inertia. In another version of the rate gyro, the reference frame, in place of being rigidly attached to the spacecraft, is rotated by an electric motor so as to keep ~o2,hence ~, near zero at all times. In this case 0 serves as the error signal in a feedback control loop that controls the rotation of the reference flame. The spacecraft rate of rotation ~2 can then be obtained directly as the negative of the rate of rotation of the reference flame relative to the spacecraft. The precision of this type of high-quality rate gyros is of the order of 0.01 deg/h. The output of an i n t e g r a t i n g gyro is the time integral of the angular rate, hence the angle through which the spacecraft has rotated in a given time interval. Integrating gyros are often the primary m e a n s for attitude determination. Because of the inevitable small drifts of such gyros, their output needs to be updated periodically by external sensors such as horizon sensors and sun sensors. Distinct from rate gyros, integrating gyros do not use a restoring torque between reference frame and gimbal. As indicated by the h o m o g e n e o u s part of (6.38) with k = 0, disturbances are d a m p e d out exponentially. On the slow time scale of o)2(t), with a sufficiently high d a m p i n g coefficient, the significant solution of (6.38) is the pseudostationary one given by o(t) =
Iav [ ~o2(r) dr + const. C J
(6.41)
In this case too, 0 can be kept small by rotating the reference frame, controlled by a closed loop where 0 serves as the error signal. The angle through which the spacecraft has rotated in a given time interval about the gyro input axis is then the negative of the readout of the shaft encoder on the motor that rotates the reference frame.
6.4.5
Inertial Measurement Units Inertial m e a s u r e m e n t units consist of a platform together with gyros and accelerometers on it. The platform is usually attached to the vehicle by gimbals. Most frequently there are three gyros and three accelerometers that have mutually orthogonal sensing axes, two parallel to the reference plane of the platform and one perpendicular to it. In a frequently used arrangement, the output from the gyros is used to maintain the platform by servo controls in a constant orientation relative to inertial space. The platform's gimbal angles then define the Euler angles of the spacecraft reference axes with respect to inertial axes. The accelerometers provide a m e a s u r e m e n t of the three c o m p o n e n t s of the resultant of the external forces other than the gravitational force that act on the spacecraft. (Gravity, of course, acts on the accelerometer's sensing masses just as it does on the vehicle.) Making use of the estimated
240
C H A PT E R 6
Attitude Control
position of the vehicle, the gravitational force can be c o m p u t e d and added to the force derived from the accelerometers. When integrated, this will give the vehicle's velocity in inertial space. The accuracy of the type of accelerometers used in space systems is of the order of 10 -6 times normal gravity, or better. In orbiting spacecraft, particularly earth-pointing spacecraft on circular orbits, it is sometimes preferred to maintain the platform orientation such that it stays locally horizontal, instead of being constant in inertial space. In this case, gravitation does not add to the acceleration along the two axes that are parallel to the platform. The gimbals are torqued to maintain the platform at zero roll and yaw. The pitch rate is controlled to be the negative of the orbit rate (27r divided by the orbital period). Inertial m e a s u r e m e n t units have found applications in spacecraft but are particularly suited for controlling launch uehicles, where they have replaced radio guidance. For a launch in the earth's atmosphere the attitude control cannot reliably use sun, star, or horizon sensors, which would be sensitive to weather and to day versus night conditions. By sensing the three c o m p o n e n t s of the acceleration and using dead reckoning, inertial m e a s u r e m e n t units can control the path of launch vehicles to high precision. On spacecraft, gyro drift affecting attitude control can be corrected by periodic updating, which is available from horizon, sun, or star sensors. In the long term, accelerometers introduce appreciable velocity and position errors. The primary m e a n s for correcting these is by periodic measurements of the Doppler shift of microwave signals emitted from the spacecraft. A particular type of inertial m e a s u r e m e n t units is referred to as strapdown IMU. As the n a m e implies, the platform is attached rigidly to the vehicle. The outputs from rate gyros are integrated by computation to obtain the Euler angles between spacecraft-fixed axes and inertial axes. Similarly, the output from the accelerometers serves to c o m p u t e the acceleration and incremental velocity c o m p o n e n t s along inertial axes.
6.5
Attitude Control Actuators Responding to inputs from the attitude sensors, the control system comm a n d s actuators, which then cause the spacecraft to reorient itself from a perturbed attitude back to the nominal one. A complete attitude system comprises both sensors and actuators. It is sometimes necessary to distinguish between 1) attitude changes attributable to external torques such as those produced by attitude control thrusters and perturbations illustrated in Fig. 6.1, and 2) internal torques produced by spacecraft internal actuators such as "reaction wheels" or caused by propellant sloshing or motions of spacecraft-internal masses. External torques change the total angular m o m e n t u m , that is, the combined angular m o m e n t u m of spacecraft body and internal moving masses. On the other hand, internal torques do not affect the total angular m o m e n t u m but can have an effect on the attitude control system because sensors
6.5 Attitude Control Actuators
241
such as horizon or sun sensors are m o u n t e d on the spacecraft body and move with it w h e n there is an internally produced disturbance. Discussed next are several types of actuators that are in c o m m o n use.
6.5.1
Thrusters Attitude control thrusters, sometimes also referred to as "reaction control thrusters," can be of two types. On some spacecraft relatively large thrusters are needed to produce major attitude changes in a relatively short time. An example is the reorientation of a spacecraft (and with it the thrust axis of attached rocket motors) prior to a motor firing. Thrusters are also used to control the attitude of a spacecraft throughout its life. These thrusters can be m u c h smaller because they only have to compensate for the long-term effects of external disturbance torques, which are typically very small. Most often, thrust levels of only a fraction of a newton are sufficient for this purpose. The physical characteristics of thrusters have been discussed in Section 4.6.9. The placement of thrusters on a spacecraft is strongly influenced by the vehicle's configuration. The thrust axis should have a large lever arm about the vehicle's center of mass. Mso, the rocket plume must not strike any part of the spacecraft. The n u m b e r of thrusters must be sufficient to rotate the spacecraft in either direction about three i n d e p e n d e n t axes. The control is simplified if these axes are the principal axes of the spacecraft. To rotate the spacecraft without at the same time imparting to it a change in velocity requires that two thrusters be fired at the same time in opposite directions to produce a pure couple. The same thrusters can also be used for imparting a velocity increment to the spacecraft, for example, for station keeping. To reduce the n u m b e r of propellant lines leading from the tanks to the thrusters it is advantageous to arrange the thrusters in clusters, such as the triplets shown in the diagram, Fig. 6.11. Because it is difficult to build small thrusters that may be required to fire a very large n u m b e r of times, it is often necessary to provide r e d u n d a n c y by increasing their n u m b e r beyond the m i n i m u m required. Figure 6.11 shows a hypothetical, symmetric arrangement of 24 thrusters that satisfy the several requirements just listed. Because of spacecraft configurational constraints, this particular arrangement often cannot be realized in practice. It is useful, however, for indicating how by alternatively combining the firing of selected thrusters redundancy can be achieved with a m i n i m u m of thrusters and that these thrusters can be used for both attitude control and for station keeping. Referring to the figure, a positive torque about the x~ principal axis can be p r o d u c e d by firing any one of the four pairs (A + 2) and (H - 2), (B + 2) and (G - 2), (E + 3) and (D - 3), or (F § 3) and ( C - 3). Acceleration along the positive direction ofx~ for station keeping can be produced by firing either the pair (B + 1) and (G § 1) or the pair (C § 1) and (F + 1). Similar choices also hold for the other axes. The propellant valves and gas pressurization valves usually also require redundancy, because otherwise a valve permanently stuck open or stuck
242
C H A P T E R 6 A t t i t u d e Control x3 A
f
D
f
A -3 ~ _ 1 . ~ . ~ /1_: 21 _ ~ ' ~ - 1 ~ ~ ~.2 ~
-3
C .i-2 ~
~ -1~
x1 w - - - . . ~
~
x2
I I
"e
~Center of Mass H
~r~_l.
-1 3 C -2 I
~
~j-2"
3
I- 2I~ Thruster ~m . l ~ l u m e s G
F
Figure 6.11 Hypothetical arrangement, with redundancy, of 24 attitude control and stationkeeping thrusters, xl, x2, x3; principal axes of spacecraft. closed would probably cause catastrophic impairments. A typical redundant propellant feed system is illustrated in Fig. 4.32. An attitude control feed system that would c o m m a n d the propellant flow to be proportional to d e m a n d would result in extremely small valves and thrusters and, as a consequence, would not be reliable. The fluid passages would be so narrow that clogging by impurities in the propellant would be a threat. Instead, o n - o f f controls are used where, except for a short transition period, the propellant control valves are either fully open or fully closed. The propellant control valves may stay open for as little as 10 ms, with transition times of several milliseconds. Ideally, after completion of the desired spacecraft attitude correction, there should be no residual rate of spacecraft rotation left. With on-off controls, however, there could be a (small) residual rotation as a consequence of the nonzero length of time needed for valve closure. Opposing thruster pairs may then act one against the other in succession. This p h e n o m e n o n [7] which is referred to as chatter, will necessarily result in some waste of propellant. To minimize the chatter of thrusters, a dead b a n d is introduced into the control, meaning that as long as the spacecraft motion is within a specified narrow band, no control action is taken. In the absence of further disturbances, the system will settle into a limit cycle, oscillating between the boundaries of the dead band. In a well-designed system, the period of the limit cycle will be long, thereby minimizing the propellant loss. A type of control, called Schmitt trigger, adds hysteresis to the dead band. Here, too, there is unavoidably a limit cycle, but its period is longer
6.5 Attitude Control Actuators
243
than in other on-off controls. Schmitt triggers are also used in a n u m b e r of other applications outside space technology. The theory of the Schmitt trigger assumes that the torque, say Mth, produced by a thruster pair about a specified spacecraft-fixed axis has one of the three values M0, or O, or -M0. The value assumed is m a d e to d e p e n d on the quantity
(6.42)
y + r dy/dt
where y is the angular displacement in inertial space of the axis relative to its desired, nominal orientation. (For systems that use a locally horizontal reference system, the definition of y is analogous, but it now refers to the local horizontal rather than to inertial space. The angle F may then be the pitch angle of the spacecraft, or, alternatively, the roll angle, i.e., angles that are directly inferred from horizon sensors.) The time constant r is at the option of the designer and is chosen as a compromise between the twin objectives of fast response of the system and m i n i m u m propellant expenditure. Figure 6.12 is a phase plane diagram that illustrates the operation of a Schmitt trigger. The various values assumed by Mth in different regions of the phase plane are indicated in the figure. The regions are b o u n d e d by the lines specified by y + r dy/dt-
0ll, 012, - 0 l l ,
-ol 2
Id_z
X/ M o dt
7 + 9dTIdt= -0~2-0~1 0~1 Cz2 .25 |
" / / /
.lO
Mth - + M -
o _',.i
~ - -
x.+~-
i
~ x....x
" ~ _ D e a d
Band
Hysteresis Bands
Figure 6.12 Phase plane of an on-off thruster controlwith deadband and hysteresis. Start oftrajectory in the phase plane at y - 0.10 rad, ~/I/Mo d F / d t 0.30 rad; r = 0.15~/I/M0. (Adapted from Ref. 2.)
244
C H A P T E R 6 Attitude Control The quantities Ol 1 and ~2 therefore determine the widths of the dead band and hysteresis bands. In the half-plane y + r d y / d t > a2, the thruster torque Mth = - M o tends to drive y toward the origin of the phase plane. Similarly, in the half-plane V + r d y / d t < - a 2 , Mth = +Mo drives y still more toward the origin. In the deadband, Mth = 0. In the hysteresis band, the value of Mth that occurred immediately before the trajectory entered the band persists in it. If I is the principal m o m e n t of inertia about the specified spacecraft axis
I d2y ~dr 2 = Mth
(6.43)
Therefore, with constants of integration )10 and (dy ~dr)o,
dy Mth t + ( d y ) , d---t- I -d-[ o
y -- 1Mth t2 + ( d y )
_
2 1
-~
t + )to
o
Elimination in these equations of the time then results in 1
(Mth/ I) (v - Yo) -- -~[ (dy /dt) 2 - ( d y / d t ) 2]
(6.44)
This shows that for Mth # 0 the trajectories are parabolas with axes of symmetry coinciding with the y axis. For Mth > 0 the parabolas point to the right, for Mth < 0 to the left. For Mth = 0 the trajectories are lines parallel to the y axis. In the example shown in the figure, the initial condition is such that y and d y / d t are both positive. Following a parabolic trajectory, the error angle y at first increases, then decreases. Between y + r d y / d t = ~1, and y + r dy ~dr = -a2, d y / d t = const. The trajectory then follows a new parabola up to where
y + r dy ~dr = -~1. In the absence of a new disturbance, the trajectory ends in a limit cycle b o u n d e d by the lines y + r d y / d t = i a 2 . The advantage of the Schmitt trigger is that for a given width of the d e a d b a n d the period of the limit cycle is long and the propellant expenditure correspondingly low. In theory, the limit cycle would persist indefinitely; in practice, dissipative effects such as those produced by propellant sloshing will cut it short. The choice of the constants r, al and a2 by the designer is based in part on the expected types of disturbances induced by the space environment. A compromise must be m a d e in the choice of al and u2: Small values reduce the residual, but only at the cost of a more rapidly oscillating limit cycle; conversely for large values. It is of interest to note that this attitude control system is nonlinear and nonconservative. An exchange takes place from the kinetic energy of rotation of the spacecraft to the energy that is carried offby the thruster gas. By the use of chemical rocket thrusters, residual errors in spacecraft attitude can be held to about 0.1 to 1.0 ~ This is acceptable in m a n y applications, for instance, in broadcasting satellites. Their main antennas have half-power b e a m widths that are m a n y times larger than the attitude control residual errors just mentioned.
6.5 Attitude Control Actuators
245
Electric thrusters for attitude control, such as xenon ion thrusters, do not suffer from the relative imprecision of the thrust cutoff of chemical thrusters. They allow very precise control. Their specific impulse is also much higher, an important consideration in the case of long-life spacecraft. A different and very critical problem arises when, because of a failure in the attitude control system or because of an unintentional release of pressurization gas, the spacecraft goes into an u n c o n t r o l l e d tumbling motion. The outputs from the attitude sensors, as may have been received by telemetry, are likely to be ambiguous, yet a thorough understanding of the most probable failure modes needs to be developed. Commanding a corrective torque about one axis may merely cause a still larger excursion about the other axes. Therefore, to avoid excessive propellant waste by the recovery action, a computer simulation, based on the full Euler's equations and the consequences of the proposed recovery action, is an essential task to be carried out before commanding the spacecraft.
6.5.2
Reaction Wheels and Gimbaled Momentum Wheels Among the devices for attitude control, other than thrusters, are reaction wheels (Fig. 6.13). The wheels are driven by electric motors in either direction and are capable of high rotational speeds. They are supported by
Figure 6.13 Schematic of reaction wheels and magnetic desaturation solenoids. (a) Assembly of three orthogonal wheels and their desaturation coils; (b) rodlike desaturation solenoid with ferrite core. L, angular momentum of spacecraft and wheels; Bg, earth magnetic field.
246
C H A P T E R 6 A t t i t u d e Control bearings that are fixed to the spacecraft. In what follows, it will be assumed that the axes of the wheels are parallel to the spacecraft principal axes. Reaction wheels serve to transfer angular m o m e n t u m from the spacecraft to the wheels and vice versa. The current, and therefore the torque, of the motors is controlled with the aim of minimizing the spacecraft attitude errors. DC motors are used because they allow precise proportional control. The accuracy of the attitude control that can be achieved is superior to what can be obtained by thrusters alone. The disturbance torques over time may change direction so that the accumulated m o m e n t u m of any of the wheels may either increase or decrease. In general, however, sooner or later the rotational speed will reach the structural limit of the wheel. This then requires periodic interventions. They are referred to as desaturations and are intended to bring the rotational speed back to near zero. Desaturation is accomplished by applying an external torque to the spacecraft. There are several methods available to do this: One approach is to fire an appropriate set of thrusters for the appropriate length of time. In orbiting spacecraft the burning of propellant can sometimes be avoided by making use of the gravity gradient. Still a third method, discussed in the next section, is magnetic desaturation. Figure 6.13 shows schematically three orthogonal pairs of reaction wheels parallel to the spacecraft principal axes X1, X2, X3. In some cases only a single wheel is used. This may be the case with Earth- or Mars-orbiting spacecraft at low altitudes, where aerodynamic drag becomes significant. Because of the asymmetry introduced by the antenna dishes, feeds, and other appendages, the aerodynamic torque will be predominantly along the pitch axis. A reaction wheel aligned with this axis can then provide accurate control. A similar situation occurs when descending toward a planet by aerobraking. Here, too, the aerodynamic torque will predominantly be about only one of the spacecraft axes. Let w be the angular velocity of the spacecraft about this axis, which is assumed to be a principal one, and let I be the corresponding moment of inertia of the spacecraft (other than the reaction wheel). The other components of the angular velocity are at first assumed to be zero. Correspondingly, let ] be the wheel's m o m e n t of inertia about its axis and ~o + v its angular velocity, where v = v(t) is the rate of rotation relative to the vehicle. The centers of mass of spacecraft and wheel are assumed to coincide. The combined angular m o m e n t u m about the center of mass is therefore L = I w + J (w + v).
The equation of motion for the spacecraft is I d) = -Nel + cv + Md
(6.45a)
] (& + i~) = +Nel - cv
(6.45b)
and for the reaction wheel
where Nel is the torque exerted on the wheel by the electric motor, Md the external disturbance torque acting on the spacecraft, and c the combined friction coefficient of the bearings.
6.5 Attitude Control Actuators
247
The armature voltage, V, that is applied to the m o t o r is controlled so as to bring the spacecraft attitude back to the desired n o m i n a l value. There will be an i n d u c e d voltage, or " b a c k - E M E " say E, and a resistive drop i R (i = armature current, R = resistance). The i n d u c e d voltage is directly related to the m o t o r torque by (6.46)
i E = gelv
(The equation follows i m m e d i a t e l y from conservation of energy of an ideal, lossless motor.) The m o t o r torque per unit current, qi - Nel/i is a c o n s t a n t for a given motor. The a r m a t u r e voltage can therefore be expressed by V = iR + E
= gel
(6.47)
R / q i + qiv
Even in the absence of an external torque on the spacecraft, there will be a small m o t o r torque and current caused by bearing friction. The resulting a r m a t u r e voltage, say, V0, o b t a i n e d from setting Md = o3 = ~ = 0, in (6.45) a n d making use of (6.47), is Vo = ( c R / q i + qi)v
(6.48)
so that V-
(6.49)
Vo - (R/qi)(Nel - cv)
Solving for Nel and substituting into (6.45a) results in the basic e q u a t i o n that characterizes the system: I
d 2t~ dt 2
=
qi R
(V-
Vo) + Md
(6.50)
Here, o5 has b e e n replaced by the second derivative of the spacecraft attitude angle relative to the inertial reference [noting that by virtue of (1.1), & = doJ/dt]. The a r m a t u r e voltage V is measured. Also measured, by a tachometer, is the angular velocity difference v. From (6.48) a n d for given m o t o r properties c, R, and qi, one can obtain V0 as a small correction to V. The wheel's m o m e n t of inertia enters (6.50) only t h r o u g h the a r m a t u r e voltage, for from (6.47) and (6.45b), V
Rl(&+i~)+ qi
qi+
v
The inputs to the control are the deviation, O, from the n o m i n a l vehicle attitude, and its derivative. The o u t p u t is the a r m a t u r e voltage, V, applied to the motor. We a s s u m e the control law V-
k(t~ + r dO~dr)
(6.51)
where the factor k and the time c o n s t a n t r are c h o s e n to ensure stability yet rapid response. The control diagram, Fig. 6.14, shows the system with its interacting c o m p o n e n t s . The p e r f o r m a n c e can be illustrated by a s s u m i n g that after s o m e instant, say tl, external disturbances are absent. The bearing friction will be neglected, hence V0 -- 0. The attitude deviation a n d its derivative in inertial
248
C HA PT E R 6
Attitude Control
J Tachometer I
Oref
"-
Control
Md
+
Reaction --- Wheel
Gyro dO/dt
Feedback
Figure 6.14
Diagram of a reaction wheel control.
space at time t - tl are designated by Ol and t~, constituting the initial condition. The trim solution t9 = a exp(st), where s is the (complex) frequency and a the amplitude, yields as the characteristic equation for s the two solutions
s - q i k ~2IR I 1+ ~/ 1 qikr 4IR 2 J
(6.52)
If and only if k > 0, the two solutions Sl and s2 are both in the negative real half-plane, indicating stability. For critical damping,
k - 4[R/ (qi'c2) The system is seen to follow in the phase plane the path given by
O(t) - al exp(sl ( t - tl)) + a2 exp(s2(t- tl)) O'(t) - alsl e x p ( s i ( t - tl)) + a2s2 exp(s2(t-
| tl))
/
(6.53)
where al
= -
S2 --Sl
,
a2
--
S2 --Sl
Concerning the technical implementation of reaction wheels, critical elements in their design are the ball or roller bearings that support them in the vehicle. The wheels must operate at high speed for long periods of time, which sometimes must exceed 10 years. Even though the bearing loads in the near-weightless environment of the spacecraft are low, wear will inevitably affect the bearing friction. Nevertheless, precision bearings with space-qualified lubricant, when thoroughly tested for qualification and flight acceptance, have proved to be a satisfactory solution. Bearing wear can be avoided if the wheels are levitated magnetically, as illustrated in Fig. 6.15. The levitation of a rotor requires the control of five degrees of freedom, three for the displacement of the center of mass and two for the tilting of the axis. In this design, a samarium-cobalt permanent magnet is used for simultaneously stabilizing against a displacement along the axis of rotation and against tilting. The remaining two degrees of freedom are the lateral displacements, which are stabilized by solenoids driven by electronic controls in response to signals from magnetic pickups. Unavoidably, there will be some eddy current losses, resulting in a small drag. The coefficient c, introduced earlier, is therefore not strictly zero.
6.5 Attitude Control Actuators
249
Figure 6.1S Schematic of reaction wheel with magnetic suspension. Courtesy of TELDIX GmbH, Germany. Magnetically suspended reaction wheels in addition still require mechanical bearings. In normal operation there is no physical contact of the rotating parts with these bearings. They are needed, however, to avoid damage in case of a temporary power failure. They also serve to cage the wheels during launch, because the loads then are too high to be contained by the magnetic suspension. Three separate, mutually orthogonal reaction wheels (Fig. 6.13) can provide complete attitude control. The analysis is virtually the same as sketched before for a single disturbance torque c o m p o n e n t about a principal axis. The only new features are the gyroscopic torques present as the wheel axes are tilted. This gives rise to nonlinear terms in the Euler rigid body equations. These terms are small and can be neglected because the wheels are operated at rates of rotation m u c h larger than any expected angular velocity of the spacecraft. Whereas the orientation of reaction wheels is fixed relative to the spacecraft, gimbaled m o m e n t u m wheels are supported by a gimbaled platform. Torques from electric motors ("torquers") are applied to the gimbals so as to tilt the wheel axes relative to the spacecraft. Figure 6.16 illustrates a type. The reactions from these torques act back on the spacecraft, returning it to its nominal attitude. The vehicle's angular m o m e n t u m is therefore transferred to the angular m o m e n t a of the wheels. In principle, a single gimbaled m o m e n t u m wheel will suffice for complete three-axis control. More c o m m o n are combinations of three wheels with orthogonal axes. An advantage is that they can provide r e d u n d a n c y in case of wheel failure. Like reaction wheels, gimbaled m o m e n t u m wheels may require periodic desaturation. The accuracy that can be achieved with either reaction wheels or gimbaled m o m e n t u m wheels is of the order of 1 microradian or better, m u c h higher than is possible with thrusters. An important application is to spacecraft-to-spacecraft and ground-spacecraft-ground laser communication. The laser pointing, acquisition, and tracking requirements are of this order. It is sufficient that the pedestal that carries the lasers, rather than the entire spacecraft, be controlled to this accuracy by reaction or m o m e n t u m wheels.
250
CHAPTER 6 Attitude Control I
Torquer
1
OuterGimbal ~----J~" InnerG.~imb~~~~. Rotor ( "~ ',~~1) .~j~.~ ~
Figure 6.16 6.5.3
.~'~
MountingPlane
~._.~.Torquer
Schematic of a gimbaled momentum wheel.
Magnetic Desaturation Desaturation consists of applying to the spacecraft an external torque opposite in direction to the wheel spin. By conservation of angular momentum, the wheel's rate of rotation will decrease, forced by the electric motor or torquer. By firing thrusters, or by a torque derived from a planetary magnetic field or gravity gradient, the wheel's rate of rotation can be returned to near zero or even reversed. This procedure is initiated autonomously by the spacecraft computer or can also be c o m m a n d e d by ground control. The discussion here is limited to magnetic desaturation. Magnetic desaturation is feasible for low-earth-orbiting satellites and spacecraft that orbit other planets with an appreciable magnetic field. Near the earth, the magnetic field can be represented approximately by a dipole field. The axis presently intersects the earth surface at about 78~ 69~ near Thule in Greenland, and at 78~ - 111 ~ near the Vostok Station in Antarctica. Based on the dipole approximation, the two magnetic poles and a magnetic equator can be defined. In terms of the spherical coordinates r, 0, q~(r = radius; 0 = geomagnetic colatitude, i.e., latitude angle measured from the magnetic pole in the northern hemisphere; q~ = geomagnetic longitude) the components of the dipole field are (e.g., Ref. 8) Br -
2Dm r3 cos 0,
B0 -
Dm r3 sin 0,
B~ = 0
(6.54)
where Dm is the magnetic dipole moment. At present Dm, in SI units, is approximately 8.1015 (tesla. m3). It follows that on the earth's surface, on the geomagnetic equator, Be = - D m / r ~ - - 3 . 1 . 1 0 -5 tesla (or in cgs units -0.31 gauss). At the magnetic poles, the field strength is twice this value. The field lines (outside the earth's iron-nickel core) run from south to north.
6.6 Spin-StabilizedVehicles
251
The dipole approximation fails at distances near and beyond the nominal position of the m a g n e t o p a u s e and even at shorter distances during magnetic storms. Magnetic desaturation is not practical at the altitude of geostationary satellites, because the magnetic field, which falls off with the third power of the distance, is too weak and too irregular there. Solenoids on the spacecraft are often arranged in pairs for each axis, as shown in Fig. 6.13. If each solenoid is approximated as a planar current ring of circular area A, the torque, M m , exerted on the pair of rings by the planetary magnetic field is Mm
--
ni(A x B)
(6.55)
where i is the current and n the total n u m b e r of turns for the pair. The vector A, with magnitude equal to the area, is directed along the solenoid axis in the direction given by the right-hand rule. Carrying out the vector product results in Mm, r
-
Mm,r
--
niDmr-3Ar sinO, M m , o - - -2niDmr-3Ar c o s O ] -niDmr-3(2Ao cos 0 - Ar sin 0) j
(6.56)
For instance, for a 400 km altitude orbit over the magnetic poles, going from south to north, and a pair of solenoids with i = 1 amp, n = 2000, A = 0.50 m a, the torque exerted on the spacecraft as it passes the magnetic equator is
Mm,~ = niDmr-3A = 0.025 (Nm) If the current flows in the direction such that A is pointing vertically downward, there will be a pitch-up torque on the spacecraft. This m a g n i t u d e of 0.025 (N m) is about 1000 times the combined torques of solar radiation, magnetic, and gravity gradient at this altitude. Spacecraft with magnetic desaturation are equipped with three magnetometers having mutually orthogonal sensing axes. The m a g n e t o m e t e r s serve to measure the strength and direction of the planetary magnetic field prior to the initiation of the desaturation. In contrast to gravity gradient desaturation (which cannot desaturate the yaw control), magnetic desaturation can be used to desaturate not only pitch and roll but also yaw. Depending on the orbit, suitable spacecraft locations must be chosen. For instance, for a polar orbit, this m a y be a location not too far from a magnetic pole (which allows pitch and roll desaturation) or from the magnetic equator (which allows pitch and yaw desaturation).
6.6
Spin-Stabilized Vehicles A useful distinction a m o n g spin-stabilized vehicles can be m a d e b e t w e e n
spin-stabilized vehicles in the more narrow sense and dual-spin vehicles. In the first case, the entire spacecraft is spinning a b o u t some axis; in the second case only a portion (usually the more massive one) does so. In a still wider sense of the term, momentum-biased spacecraft, that is, spacecraft that incorporate a fast spinning m o m e n t u m wheel (Sect. 6.5.2), can also be classified as spin stabilized. In all cases, the orientation in inertial space of
252
C H A P T E R 6 Attitude Control the spin axis is maintained by the gyroscopic effect that is produced by the spinning part. The attitude control of spin-stabilized spacecraft (Fig. 6.2a) is simpler than it is for three-axis stabilized vehicles. This advantage can be offset, however, bythe dimensional restrictions that are imposed bythe available launch vehicle payload space (i.e., the size of the "shroud" or "fairing"). Because the m a x i m u m possible solar cell area of spinning vehicles is proportional to their diameter and length, high-power communications and broadcasting satellites with their high d e m a n d for electric power, w h e n spin-stabilized, can become too large for available launch vehicles. Spinning spacecraft can be either prolate (Fig. 6.4a) or oblate (Fig. 6.4b). As discussed in Sect. 6.3.2, the precession ofprolate bodies is unstable in the sense that the nutation angle tends to increase. Oblate bodies are stable in this sense. However, the attitude control of spinning vehicles requires more than merely a stable precession at a constant nutation angle: a control is needed to reduce to near zero the nutation angle that may have been caused by a disturbance. In spite of their basic instability (which can be corrected by nutation dampers; see Sect. 6.6.3), prolate spacecraft are often preferred because of their better geometrical fit to the launch vehicle payload space. Horizon sensors can take advantage of the spin. Mounted on the side of the vehicle, the line of s i g h t - - by virtue of the spin m can be made to sweep over the earth or planet (Fig. 6.17). If a stepping motor is used in addition to change the line of sight in the fore-aft direction, a single sensor can serve for the sweeps A to A' and B to B' indicated in Fig. 6.8.
6.6.1
Jet Damping Rocket motor gas, as it streams through the motor case and the nozzle, generally has a d a m p i n g effect on spacecraft motions. The p h e n o m e n o n is referred to as jet d a m p i n g [9]. As in aircraft, it can d a m p e n pitch and yaw J (
~
J . " Damping " Wheel
X3
Horizon~ Sensor
/.
/
~,
Stepping
Motor Rotation
x
i
/
Line , of Sight
Figure 6.17 Schematic of a spinning spacecraft with nutation damping wheel and horizon sensor (wheel axis parallel to xl principal axis).
6.6 Spin-Stabilized Vehicles
253
motions. More significant, however, is the jet damping of s p i n n i n g vehicles, because it counteracts the precessional motion by decreasing the nutation angle. The beneficial effect ofjet d a m p i n g obtained during rocket motor firings is important because it mitigates the relatively large disturbance torques that are caused by unavoidable motor performance fluctuations. Nevertheless, by itself, jet d a m p i n g may not be sufficient to keep the nutation angle from growing during motor firings. There have been some upper stage vehicles with solid-propellant motors that exhibited unacceptably large nutation angles. To keep the angle within one or two degrees then required the addition of special thrusters [10-12]. Jet damping can be explained by the torque that is produced by the Coriolis forces resulting from the angular velocity of the spacecraft and the rocket gas velocity relative to it. The jet damping torque is the integrated m o m e n t of these Coriolis forces, starting at the combustion c h a m b e r and ending at the nozzle exit. An explanation that is both simpler and also makes it clear that the combustion c h a m b e r and the nozzle are not separately involved, is as follows. A simple but sufficiently accurate model of jet d a m p i n g is illustrated in Fig. 6.18. The model assumes that the mass-averaged gas exit velocity, Uex, at the nozzle exit and relative to the spacecraft is in the direction of the nominal thrust. (This is not quite accurate because the Coriolis forces cause the gas stream at the nozzle exit to deviate slightly from this direction. The gas already enters the nozzle with some angle of attack. But because x3
o
8
Mij
r1
Mean Burn Surfac
ul
-
Rocket Motor
Nozzle u2
~lUex
Figure 6.18 Jet damping of a spinning spacecraft: O, instantaneous center of mass; O, nutation angle; p, precession rate; Uex, gas exit velocity relative to spacecraft; Mid, jet damping moment. Shown for case of prolate spacecraft.
254
C H A P T E R 6 Attitude Control
of the subsequent large acceleration in the nozzle, the effect of the angle of attack is largely canceled and the gas stream is redirected to leave the nozzle close to the nominal thrust direction. The most thorough investigation, both theoretically and experimentally, of nozzle flows with initial angle of attack has been published by Pirumov and Roslyakov [13].) The figure refers to a spinning, inertially symmetric, prolate vehicle with a solid-propellant motor, but the results derived in the following are equally valid in the oblate case. Let x3 designate the spacecraft principal axis that coincides with the nominal thrust axis. Also, let rl be the position vector in the aft direction of x3 from the (instantaneous) center of mass, O, to the average location of the burn surface at a specified time. Similarly, r2 is the position vector from O to the nozzle exit midpoint. The spacecraft angular velocity is designated by ,~, the m o m e n t u m vector by L, the nutation angle by 0, the precession rate by p, and the gas mass flow rate by ~. A onedimensional, steady-state description is used. Because the damping rate is slow compared with w, it suffices to compute the jet damping m o m e n t separately at each instance, assuming that all quantities are constant at that time. As a consequence of the precession, there is at the nozzle exit, relative to inertial space, an additional velocity c o m p o n e n t of the gas u2 = p (L/L ) x r2
transverse to the thrust axis. The reactive force resulting from the momentum per unit time, r~fi2, is opposite to the direction of the precession and tends to decrease the nutation angle. It accounts for the major part of the jet damping effect. A closed control surface can be introduced that contains the spacecraft and crosses the nozzle exit plane. The angular m o m e n t u m per unit time (in inertial space and relative to the center of mass) of the gas that leaves the control surface at the nozzle exit is r~pr2 x (L/L x r2)
(6.57a)
The angular m o m e n t u m loss in the control volume, per unit time, caused by the propellant diminution at the burn surface is -~prl
x (L/L x rl)
(6.57b)
The torque on the spacecraft, balancing these two angular m o m e n t u m changes, is the jet d a m p i n g m o m e n t
Mid =/f/p[r2 x (L/L x r 2 ) - rl x (L/L x rl)]
(6.58)
The vector Mid is seen to be perpendicular to the thrust axis and to rotate around it. The second term in the square bracket is generally m u c h smaller than the first, since the center of mass of most spacecraft is closer to the burn surface than to the nozzle exit plane. Therefore, the averaging of the location of the burn surface, as was done in defining rl, will not generally introduce appreciable errors.
6.6 Spin-Stabilized Vehicles
255
It follows that the magnitude Mid of the jet damping m o m e n t is Mid = rhp(4
-
rl2) sin 0
-
rhino3 (r22 - r 2) tan 11
0
Therefore, for small nutation angles _
13 iF2
]~dj d - - m--~l W t 2 --
6.6.2
rf ) O
(6.58')
Passive Nutation Damping of Oblate Spacecraft Spinning, inertially symmetric spacecraft that are o b l a t e , after disturbances have subsided, will continue to precess at a constant nutation angle (Sect. 6.3.1). In this sense, they are stable. The need, however, is to reduce the nutation angle to near zero after a disturbance. A n u m b e r of different passive nutation dampers are in use. The principle upon which they are based is the same: The precessional motion of the spacecraft induces a motion of the damping element relative to the spacecraft, thereby dissipating energy. The arrangement shown in Fig. 6.17 can serve as an example: A small wheel is m o u n t e d with its axis of rotation perpendicular to the vehicle's nominal spin axis x3. The wheel is passive in the sense that it is not driven by a motor but merely responds to the spacecraft motion. Provision is made for damping the wheel's rotation (which can be in either direction), typically by eddy currents induced by a magnetic field. Let w(t) be the angular velocity of the spacecraft body relative to inertim space, v(t) the rate of rotation of the wheel, and ] the axial m o m e n t of inertia of the wheel. The x~ principal axis is taken parallel to the wheel's axis. With il, i2, i3 designating the unit base vectors of the principal axes coordinate system, the combined angular m o m e n t u m , L, of spacecraft body and wheel is L = (Ilwl + ] v)il + I20~2i2 +/3w3i3 Let Md be the disturbance torque that acts on the spacecraft. Then Md = d L / d t = L + o: x L where, as before, (') indicates time derivatives in the spacecraft-fixed space a n d d ( ) / d t in inertial space. Substituting for L, and noting that the basis vectors and the m o m e n t s of inertia are constant in the dot differentiation, it follows that I1(-bl + ]1) + (% -- 11)o)2o93 = Cv + Md,1
(6.59a)
/16)2 - (% -- 11)o)1o)3 + ]w3v = Md,2
(6.59b)
I30)3 -- ](.02V -- Md,3
(6.59C)
where c is the damping coefficient, hence the torque cv the reaction from
256
C H A P T E R 6 Attitude Control the d a m p e d wheel on the spacecraft. The wheel satisfies the condition (6.60)
] ((J)l "['- 1)) = --CU
The transverse angular velocity 092 will be m u c h smaller than the vehicle's rate of spin. It follows from the last of equations (6.59) that 093, substituted into the first two, can be approximated by a constant, say 090. These two equations, together with (6.60), then result in the set of linear equations represented by
[110,1I 11 [ 0 0
11 0
]
0
]
(J)2
+
i 0ii, o
- (13 -- I1) 090
~)
0
] 090
092
C
V
0
--
[Mdx1 Md, 2 Md,3
(6.61) To examine the stability of the system and its rate of approach to equilibrium after an external disturbance has subsided, it suffices to examine the h o m o g e n e o u s part of (6.61). Making the substitutions 0)1 m
al eSt,
a2e st,
092 - -
v = ave st
where s is the (complex) frequency and aa, a2, a~ are the (complex) amplitudes, the existence of nontrivial solutions therefore requires that Il s
(13 -- 11)o90
l S -- C
-(13 - 11)09o
Il s 0
109o Is + c
]s
-
0
(6.62)
If the root locus method is applied for examining the stability of the motion, it becomes convenient to cast this equation into Evan's form, which makes evident the zeros and poles in the complex plane of s. In this form, (6.62) becomes s(s 2 + Z2092) S2 +
(1 + l / I 1 ) c
(6.63>
(1 -- ] / 1 1 ) ]
(I3/I' -1)2w2 1+1/11
where ~ 2 __
( I 3 / I ~ - 1)(13/11 + ] / 1 1 1 -
1)
]/11
It is found that, with suitable choices of ] and c, this passive means of stabilizing is adequate to reduce the nutation of o b l a t e vehicles to zero after a disturbance. For p r o l a t e v e h i c l e s , either an active system (e.g., with wheels that are driven by electric motors in response to gyro inputs) or the passive system described in the next section is usually required. An example is shown in Fig. 6.19 for m o m e n t ofinertia ratios 13/I1 = 1.50 (an oblate vehicle) and ]/I1 = 0.025. Shown as functions of the nondimensional damping coefficient (defined in the figure) are the loci of the three roots in the complex plane ofs. There is one pair of conjugate complex roots and one purely real root. The latter is highly damped. The critical ones are the complex ones, because they result in the longest time for the exponential
6.6 Spin-Stabilized Vehicles | 'C(1 + J/I 1)
.52 I-" ~OJ(1-J/I1) = .51
|
'
4o ' 3 o . s
.50 I 3>
.20
.60~..~ 1.00
7 _o o1_~
_1
.10 _ 2.60
257
O-1
-
"
Ol
o~
roots
1.00 .60~,f~ ~
/
-50 1
-~'~ -~I / I
2.00 -"
~o~ .,~ .~o .~o--T~~~-o
-.015
oI
-.010 -.005 Re(s/e%) I
0
I
Figure 6.19 Passive nutation damping of oblate spacecraft: root loci for 13/11 - 1.50 and l/I1 = 0.025. From Eq. (6.63). decay of a disturbance. The o p t i m u m d a m p i n g coefficient is approximately 0.53 (where the t a n g e n t to the curves is vertical). Figure 6.20 shows as a function of o~0t the d a m p e d oscillations of a vehicle with the same ratio 13/11 = 1.50. The angular velocities o~, ~o2 a n d the n u t a t i o n angle 0 are p r o p o r t i o n a l to the ordinate shown. Represented are the least d a m p e d m o d e s for each of three values of ] / h , each with its optimal d a m p i n g coefficient. As is to be expected, the larger the wheel's 1/
.
.
.
.
.
.
.
.
|
FJ/I1 = 0.025
A
~
0.100
q'._ 02 o~
ATL~
"0
-1o
~'o
~o
~'o
~'o
1
t
1oo 1~o 1~o 1~o
mot
Figure 6.20 Passive nutation damping of oblate spacecraft: least damped modes with optimal damping. For 13/11 - 1.50 and three values of 1/11. From Eq. (6.63).
258
CHAPTER
6 Attitude Control
m o m e n t of inertia, the more rapid is the decay from an external disturbance.
6.6.3
Dual-Spin Spacecraft Orbiting, spin-stabilized vehicles usually require a despun platform for the support of antennas, sensors, imagers, or scientific instruments that are intended to point toward the earth or a planet. The rate of rotation of the platform is therefore comparable to the orbital rate, which differs from and is m u c h smaller than the vehicle spin. These vehicles are often called dualspin. A typical configuration of spin-stabilized vehicles for space c o m m u n i ' cations or earth observations is shown in Fig. 6.2a. The despun platform and rotor are connected by a shaft supported by bearings in the rotor. The torque from a small electric motor compensates for the bearing friction. This motor is controlled by an error signal, derived, for instance, from horizon sensors that detect deviations from the intended orientation of the platform. Integral parts of the bearing assembly are slip rings that provide the needed electrical connections between platform and rotor for power and signal circuits. Usually, only low-frequency signals are transmitted in this way. Microwave frequencies are usually generated directly on the platform, to avoid rotating waveguide joints. The larger part of the spacecraft mass resides in the rotor, which carries the solar cells, power conditioning equipment, batteries, spacecraft propellant, and frequently also a solid-propellant motor as the last stage in the flight trajectory. Concentrating the mass in the rotor is desirable for it enhances the short-term attitude stability. The rate of rotor spin may be quite low (e.g., one revolution per second, which would still make it about 5000 times larger than the platform rate of a low-altitude earth-orbiting satellite). Active damping, for instance by wheels similar to the wheel shown in Fig. 6.17, but driven, can bring the nutation angle back to zero, even in the case of prolate vehicles. However, it was discovered [14], that the same effect can also be obtained by purely passive means, provided that a n u t a t i o n d a m p e r is located on the despun platform. An example of such a damper is shown schematically in Fig. 6.21. Other arrangements, such as a damped pendulum, can be equally effective. For the purpose of analyzing the stabilizing effect of the nutation damper that is illustrated in the figure, it will be assumed that both the rotor and the despun platform are inertially symmetric about their axis of relative rotation, x3. The combined center of mass, O, of rotor and despun section is located on this axis and is taken as the origin of the spacecraft's principal axes Xl, X2, X3. These are taken as corotating with the despun platform. The vectors ofthe coordinate system formed bythe principal axes are designated by i l, i2, i3. The rotor's angular velocity relative to inertial space is ~ = ~lil + ~2i2 + ~3i3 and its m o m e n t s of inertia about O are I1 = 12 and 13. The corresponding data for the despun section are f~ = ~lil + ~2i2 + g23i3, fl = ] 2
259
6.6 Spin-Stabilized Vehicles
rn
/~ 2
x•
r~ 2 x1
r of Mass
X3
o,o, x3
~'/2 = 0)2
k
~-- Rotor (co) Despun Platform (f2) Nutation Damper (a)
(b)
Figure 6.21 Nutation damper on dual-spin spacecraft: (a) schematic, (b) derivation of Eq. (6.67). (Xl, x2, x3) -- plateform fixed principal axes; ,~ = rotor angular velocity; fl = plateform angular velocity. For clarity, the length of the nutation damper relative to the spacecraft is greatly exaggerated. and 13. By virtue of the constraint imposed by the c o m m o n axis of rotation, ~'~1 "-- 0)1,
(6.64)
~"~2 = 032
As shown in the figure, a mass, m, is free to move inside a tube located on the despun section, through x3 and parallel to xl. The mass is acted on by a pair of springs with combined spring constant k and by a damping force proportional to the relative velocity of the mass and the tube, with damping coefficient c. The distance of the tube from the center of mass, O, is designated by I and the distance of m from the x3 axis by y(t), where, by assumption, lYl ~ l. To be effective as a damper, k and c are chosen such that the springmass system is approximately in resonance with the vehicle's precession frequency [given by (6.21)]. The latter, to second-order accuracy for small nutation angles, is a constant for a given spacecraft with fixed m o m e n t s of inertia and rate of spin of the rotor. That the nutation angle is assumed to be small implies that If~l i, 1~21, 1fl3i ~ coo, where oJ0 is the rotor's nominal rate of spin, a constant. Time derivatives in the space fixed to the despun section will be designated by ('), those in inertial space by d( )dr. To estimate the orders of magnitude of the various terms, it is convenient to write ~ 1 = wOel,
~ 2 -- WOe2,
f23 :
wOe3,
y = lo
where ISll, Is21, le3i, 1,71 << 1. Since the rate of precession o f t h e vehicle is of the same order as wo, each (') derivative introduces a factor of wo into the estimated order o f m a g n i t u d e ofthe terms. Thus, for instance, ~2~ -- O(w2E~) and Y ~ 2 - - O(lw21?e2).
260
C H A P T E R 6 Attitude Control Considering next the m o t i o n of the mass m, its position vector from the vehicle center of mass, O, will be designated by r(t). Therefore
r-
yil + li3,
i"- J)il,
i ~ - j)il
If F is the resultant force exerted on m by the spring, the viscous tube friction, and the lateral force from the tube, then m ~ - F - m[f~ x (~2 x r) + 2f~ x f + ~ x r]
(6.65)
where the term with the square bracket represents the inertia force, as in (1.4). (The term d2Ro/dt 2 is absent here because the vehicle is in free fall and gravity acts b o t h on the vehicle and on the mass m.) Evaluating the vector products in terms of their Xl, x2, x3 c o m p o n e n t s and estimating the orders of magnitudes, one finds that all terms in the square bracket are of second or third order in e and 0, except for two first-order terms that arise from the fourth vector product,
x r - / E 2 2 i l -/~21i2 Dropping all terms of higher order t h a n the first, (6.65) b e c o m e s F = (m~ + m/O2)il - m/fall2
(6.66)
The force exerted by m on the vehicle is - F , which results in a torque, say Nm, on the vehicle about O, given therefore by Nm = -m/2~21il - l ( m ) + rn/~22)i2
(6.67)
The c o m b i n e d angular m o m e n t u m , L, of rotor, d e s p u n platform, a n d d a m p e r mass is L-
(11 +/1)g21il + [(I1 + 11)g22 + m/y]i2 + (I30)3 -[- 13~3)i3
(6.68)
From conservation of m o m e n t u m d L / d t - Nm + Md = (--m12~21 + Mo,1)|I + ( - m l ) -
m12~2 + Md,2)i2 + Md,3i3 (6.69)
where Md(t) is the m o m e n t of the external disturbance that acts on the vehicle and where, from (1.1) dL/dt-
L + f~ x L
Differentiating L and evaluating the c o m p o n e n t s of the vector product yields for the first a n d second c o m p o n e n t s (11 + ]1 + ml2)~21 + I3w0~2 -- Md,1 (11 + 11 + m/2)f22 - I30)0f21 + 2 m l ) -
Md,2
(6.70a) (6.70b)
The third c o m p o n e n t is not needed; in its place one has the equation of m o t i o n of the d a m p e r mass, which from the first c o m p o n e n t of (6.65) is m ) + cy + ky + mlE22 - 0
(6.70c)
261
6.6 Spin-Stabilized Vehicles
For the study of the stability of the system, it suffices to consider the h o m o g e n e o u s parts. Setting ~'~1
--
ale st,
g22 = a2e st,
y = aye st
where s is the (generally complex) frequency and al, a2, ay are the (generally complex) amplitudes, a system of three linear, h o m o g e n e o u s equations for the amplitudes is obtained. The existence of a nontrivial solution requires that (Ia + 11 + ml2)s --I3o9o 0
13o90
0 (6.71)
2mls 2 = 0 m s 2 + cs + k
(h + ]a + ml2)s mls
Evaluation of the determinant results in the fourth-degree characteristic polynomial, [(/1 + l l + ml2)2m - 2(I1 + J1 + ml2)mZl2]s4 + (/1 + l l + ml2)2cs 3 (6.72)
2 2 2 = 0 + [(I1 + 11 qt_ m12)2k+ i2w2m]s2 + i3woCS + i32 o90k
The system is stable if all four roots are in the left half of the complex plane, unstable if one or more roots are in the right half-plane. If stable, the angular velocity components and the nutation angle w i l l after the external disturbance has subsided, decay toward zero, in the most general case as a superposition of exponentially damped oscillations and of pure exponentials. An example is shown in Fig. 6.22 for a spacecraftwith/1//3 = 1.00, ]~ //3 = 0.50. For the nutation damper it is assumed that ml2/I3 = 0.05 and k/ (rn~ 2) - 1. The nondimensional damping coefficient c/(nUoo) is taken as the parameter. Depending on its range, there are two pairs of conjugate complex roots or else one such pair and two real roots. An o p t i m u m damping coefficient is one that results in the most rapid damping of the least damped mode. In the present case, its value is 1.85, as is indicated in the figure, I
1.00
K
0
i
c m ~01 ~
0.50 o
I
18sI
2.00
-0.50
I
-1.20
I
-1.00
I
I
I
1.85 ~.00 ~00
10.0
1.85>~.005=.00
10.0
I
2.o0 _ -I./u - - -~...~ 9
-1.00 -
I
1.00
1.50
/
"\
%
1.50
I
-.80
1.00 /~/
I
I
I
I
I
10-3 .-1.0 -.80 -.60 -.40 -.20
0
Re(s/oo)
Figure 6.22 Dual-spin spacecraft: root loci for a spacecraft with I1/I3 = 1.00, la//3 = 0.50, ml 2//3 = 0.05, k/(mogoz) = 1. From
Eq. (6.72).
C H A P T E R 6 Attitude Control
262
6.7
Gravity Gradient Stabilization The parts of a spacecraft that are closer to the center of gravitational attraction than those at a larger distance are acted upon by a slightly larger gravitational force. Hence, in general there will be a gravitational torque. Although very small, it can be used to stabilize the spacecraft attitude by means that are purely passive. Because of the smallness of the torque, gravity gradient stabilization is useful only when the disturbance torques (Fig. 6.1) are merely a fraction of the available gravity torque. Figure 6.23 shows, as an example, an orbiting spacecraft with a long mast that extends toward the earth. The spacecraft is assumed to be in a circular orbit with radius R0 to the spacecraft center of mass. It is convenient in this case to specify the attitude by means of a locally horizontal, orthogonal system of reference axes through the center of mass. This system is familiar from aircraft practice. Thus the x; axis is in the direction of motion, the x~ axis downward toward the center of attraction, and the x~ axis such as to form with the others a right-handed system. The principal axes of the spacecraft are Xl (the "roll axis"), x2 (the "pitch axis"), and x3 (the "yaw axis"). The angles from the primed to the unprimed axes are the pitch angle 0, roll angle q~, and yaw angle ~. All are assumed to be very small. The m o m e n t s of inertia about the principal axes are 11, 12, I3. The expression for the gravitational torque has been derived in Chap. 2, Eq. (2.9). Carrying out c o m p o n e n t by c o m p o n e n t the matrix multiplication needed to obtain the tensor product ]. R0 and then the vector product R0 x (]. R0), one obtains for the first c o m p o n e n t Mg~ of the gravitational torque
Mgl
:
- 3 # R o 5 ( I 2 - I3)x~x~
~
r X1
x2 0
Circular Orbit
Mast
R~
IF
X3 X3 To Center of Attraction
Figure 6.23 Gravity gradient stabilization: spacecraft with mast in circular orbit. 0, pitch; ~, roll; ~p, yaw.
(6.73)
6.7 Gravity Gradient Stabilization
263
where tx is the gravitational parameter (IX = 3.986 105 k m 3 s 2 for the earth). The corresponding results for the other c o m p o n e n t s are obtained by cyclic interchange of the subscripts 1, 2, 3. With w designating the angular velocity of the spacecraft relative to inertial space and Md an external (nongravitational) disturbance torque, Euler's equations (6.16') for constant m o m e n t s of inertia b e c o m e Ilo91 + (/3
/2)o)2o)3= -3txRo5(I2 - I3)y~x~ + Md,]
(6.74)
for the first c o m p o n e n t and by cyclic permutation of the subscripts for the others. For small angles 0, r x;-
x~--Roe,
Ro O,
x~--Ro
Therefore the full set of Euler's equations becomes I1&1 + (/3 - I2)w2w3 - -3~R03(I2 - / 3 ) r + Mdl / I2d)2 Jr- (11 - 13)o3o91 -- +3txRo3(I3 - I1)0 + Md2 [ I30)3 + (12 -- 11)091o92 -- +Md,3
(6.75)
where a second-order term with the factor 0r has been dropped. One can introduce the orbital angular velocity, n, which is the rate at which the attitude relative to inertial space changes by virtue of the motion along the circular orbit. It follows from (3.40) that n 2 = ~/Ro3. The angular rates 0, r and ~ are assumed to be small compared with n. The orbital angular velocity causes gyroscopic effects, which are essential to the gravity gradient stabilization. For small angles 0, r ~ the angular velocity c o m p o n e n t s b e c o m e O91 -- ~
n~,
0)3 = ~ + nr
O92 -- 0 -- /2,
(6.76)
The equations governing the p i t c h are seen to be i n d e p e n d e n t of those for roll and yaw. Neglecting a second-order term with the factor wxw3, one obtains from the second of equations (6.75) I2d)2 +
3n2(I1
13)0
--
Md2
/
(6.77)
!
w2--O-n
for w2 and 0. The roll and y a w are coupled [2]. It follows from the first and the third of Eqs. (6.75), neglecting 0 in comparison with n, that I]&l § n(I2 - I3)w3 + 3n2(I2 - / 3 ) r I3d93 + n(I1
o)1 - r
12)o)1 -- M d 3
- Md,1 (6.78)
nO
o)3 = ~ + nr for Wl, w3, r and ~. The stability of the motion can be studied by limiting oneself to the h o m o g e n e o u s parts. Letting 092
--
a2
e st,
0 = ao e st
264
C H A P T E R 6 A t t i t u d e Control for the pitch equations leads to the characteristic equation I2s
1
3n2(I1 - / 3 ) -s
(6.79)
= 0
for the complex frequency s, hence to the second-degree polynomial s 2 + 3n2(I1 - 13)/12 = 0 If I1 > /3, there will be, following an external disturbance, an u n d a m p e d harmonic oscillation, referred to as "pitch libration." Its angular frequencyis r/~3(/1 - 13)
/2 If I1 < Iz, the motion is unstable with an exponential increase of the pitch angle at the rate
rt•3(I3
- I1)
/2
Letting O91 ~"
a l e st,
o93 = a3e st,
= a r e st
ck = a ~ e st,
for the roll/yaw, the characteristic equation becomes IlS n ( I 1 - 12) --1 0
/'/(/2- 13)
3 n 2 ( I 2 - /3)
I3s 0 --1
0 s n
0 0 -n s
F i g u r e 6.24 Stability of the rollyaw motions of gravity-gradient stabilized spacecraft. (Ref. 2.) Adapted from Bryson, A. E., "Control of Spacecraft and Aircraft," Copyright 9 1994 Princeton University Press. By permission.
-
0
(6.80)
6.7 Gravity Gradient Stabilization
265
leading to a fourth-degree polynomial for s. The regions for stable and unstable motions are shown in Fig. 6.24 (after Bryson [2]).
Nomenclature A B C
Dm e
E i ii l,J Ii, li k
lij L m
mc M N n
P qi r
R S
T U
v xi &j o o p T CO
dv/dt ~r
()d ()el
()ex ()jd
area magnetic induction damping/friction coefficient magnetic dipole m o m e n t eigenvector back EMF current base vectors Cartesian inertia tensors principal m o m e n t s of inertia spring constant [Eq. (6.36)]; also control gain [Eq. (6.51)] direction cosines angular m o m e n t u m spacecraft mass d a m p e r mass spacecraft external torque spacecraft internal torque orbital rate of rotation [Eq. (6.76)]; also n u m b e r of turns [Eq. (6.55)] precession rate motor torque per unit current position vector; radius resistance complex frequency kinetic energy of rotation; T: Cartesian tensor gas velocity armature voltage principal axes declination of sensor axis Kronecker delta nutation angle (Fig. 6.4); also pitch angle (Figs. 6.7, 6.23) rotor axis angle (Fig. 6.10) eigenvalue density time constant [Eq. (6.42)] angular velocity relative to inertial space time derivative of a vector v in inertial space time derivative of a vector v in body-fixed space disturbance electric exit plane jet d a m p i n g
266
C H A P T E R 6 Attitude Control
( )m ( )th
magnetic
thruster
Problems (1) In a Cartesian coordinate system with unit base vectors il, i2, i3, the inertia tensor of a rigid body is represented by the matrix (in arbitrary units)
[ ] 1
0
0
0 0
1 2
2 1
(a) Find the eigenvalues. Also find the directions of the principal axes of inertia. (Choose the algebraic signs ofthe eigenvectors such that they form a right-handed system.) Express the eigenvectors in terms of i~, |2, i3. (b) As a check on (a), verify that the principal axes of inertia, as calculated, are mutually orthogonal. (c) Determine the numerical values of the principal m o m e n t s of inertia. (2) Consider a spinning, rigid b o d y w i t h principal axes x~, Xa, x3 and corresponding m o m e n t s of inertia I~,/2,/3. The x3 axis is an axis of symmetry, so that I~ --/2. The c o m p o n e n t s of the body's angular velocity about its principal axes are designated by oJ~, o~2, ~o3. An external torque is acting on the body, with c o m p o n e n t s M1, 71//2, M3 about the principal axes. Assume that M~ = M0 = constant and 71//2 --/I//3 - 0. Initially, the angular velocity c o m p o n e n t s are a s s u m e d to be equal, say o~0. Find oJa(t) for t >_ 0. Express the result in terms of o~0, Mo, and the m o m e n t s of inertia. (3) The attitude of a spacecraft is a s s u m e d to have been disturbed impulsively about one of its principal axes. Following the disturbance, an on-off control, provided by thrusters, is attempting to return the attitude of the spacecraft back to its nominal (zero) orientation. The torque by the thrusters has constant m a g n i t u d e but can be in either direction. The m o m e n t of inertia pertaining to this axis is 1000 kg m 2. The magnitude of the thruster torque is 0.30 Nm. Let F the angular deviation of the attitude from its n o m i n a l value. Immediately following the disturbance F - 0 and d F / d t - 0.005 rad/s. A control law is a s s u m e d of the form F + ~ dF/dt
with time constant ~ = 10 s. To improve the stability, a d e a d b a n d between F = +0.015 rad is provided. There are two hysteresis zones, one between F -- 0.015 rad and 0.008 rad, the other b e t w e e n -0.008 rad and -0.015 rad (see Fig. 6.12). (a) Draw a phase diagram with coordinates F and d F / d t t h a t shows the path followed by the spacecraft attitude. (b) C o m p u t e the period of the limit cycle. (4) A spacecraft is in a circular orbit that takes it over the earth's magnetic poles. It is e q u i p p e d with a reaction wheel a n d a magnetic desaturation solenoid.
References
267
The wheel's axis is perpendicular to the orbital plane. The axis ofthe solenoid is at all times aligned with the local vertical. The orbit is at 400 km altitude (orbit radius approximately 6760 km). At this altitude the earth's magnetic field can be approximated by a dipole field with magnetic lines running from the magnetic south pole to the magnetic north pole. The dipole moment is 8 1015 tesla m 3 (= volt s m ) . The only component that needs to be taken into account is the one along the orbit. The reaction wheel has a moment of inertia of 0.02 kg m 2. At saturation it spins at 12,000 rpm. Magnetic desaturation is accomplished with a rodlike solenoid with a ferrite core (Fig. 6.13b). The number of turns is 2000, the current is 1.0 amp. The permeability (relative to the permeability of vacuum) of the ferrite core is 200. The effective cross section of the rod is 10 cm 2. The approximations valid for long, slender solenoids can be used. Compute the reaction wheel's angular momentum at saturation. If desaturation is applied over one-halfofthe orbit, from magnetic pole to magnetic pole, what fraction of the saturation angular momentum can be eliminated ?
References 1. 2. 3.
4.
Kane,T. R., Likins, P. W., Levinson, D. A., "Spacecraft Dynamics," McGraw-Hill, New York, 1983. Bryson, A. E., "Control ofSpacecraft andAircraft," Princeton University Press, Princeton, NJ, 1994. "Spacecraft Pointing and Control," North Atlantic Treaty Organization Advisory Group for Aerospace Research and Development, AGARDograph No. 260, 1982. Hughes, P. C., "Spacecraft Attitude Dynamics," John Wiley & Sons, New York, 1986.
5.
Kaplan, M. H., "Modern Spacecraft Dynamics and Control," John Wiley & Sons, New York, 1976.
6.
Wertz, J. R., ed., "Spacecraft Attitude Determination and Control," Kluvet Academic Publishers, Dordrecht, The Netherlands, 1978. Zelikin, M. I. and Borisov, V. E, "Theory of Chattering Control," translated from the Russian, Birkh~iuser, Basel, 1994. Stratton, J.A., "Electromagnetic Theory," McGraw-Hill, NewYork, 1941. Thomson, W. T. andReiter, G.S., "Jet Damping ofa Solid Rocket: Theory and Flight Results," A/AA Journal, Vol. 3, No. 3, pp. 413-416, 1965. McIntyre, J. E. and Tanner, T. M., "Fuel Slosh in a Spinning On-Axis Propellant Tank: An Eigenmode Approach," Space Communication and Broadcasting, Vol. 5, No. 4, pp. 229-251, 1987. Mingori, D. L., Halsmer, D. M., and Yam, Y., "Stability of Spinning Rockets with Internal Mass Motion," American Astronautical Society Proceedings, pp. 93-135, February 1993.
7. 8. 9. 10.
11.
268
C HAPTER 6 Attitude Control 12.
13.
14.
Meyer, R. X., "Coning Instability of Spacecraft During Periods of Thrust," Journal of Spacecraft and Rockets, Vol. 33, No. 6, pp. 781-788, 1996. Pirumov, U. G. and Roslyakov, G. S., "Gas Flow in Nozzles," Springer Series in Chemical Physics, Vol. 29, Springer Verlag, Berlin, translated from the Russian, 1986. Iorillo, A. J., "Analysis Related to the Hughes Gyrostat Systems," Hughes Aircraft Company Report 70438 B, 1967.
7 Spacecraft Thermal Design Spacecraft absorb and emit radiation. Depending on the physical characteristics of their external surfaces, they absorb more or less of the incident solar radiation. Spacecraft operating in the near-earth environment also absorb to a significant degree thermal radiation (in the infrared) that is emitted by the earth and solar radiation that is reflected by the earth (albedo effect). Analogous effects also play a role when operating near the inner planets and the earth's moon. In turn, spacecraft emit thermal radiation into space. When in thermal equilibrium, their emission equals the sum of (1) their absorption of the incident radiation of all types (in the case of solar panels their net thermal absorption plus the electric power generated by them) and (2) the heat produced by spacecraft internal sources such as the heat dissipated by electrical components, by heat transfer from the combustion ofpropellants, and sometimes by radioisotope sources. The reader will find m u c h additional information related to spacecraft radiative heat transfer in Refs. 1 to 3. Spacecraft temperatures must be controlled to within close lower and upper limits. These are imposed by the different characteristics of the various components on the spacecraft. Particularly sensitive to temperature extremes are storage batteries, spacecraft propellants, m a n y electronic components, and certain scientific payloads. The allowable limits for electronic components are not necessarily the same for operating and nonoperating conditions (when operating, the additional electric stress often imposes a lower upper temperature limit). The temperature assumed by these c o m p o n e n t s will be determined by their own heat production and, on the other hand, by the transfer of heat by thermal conduction and radiation to and from other components and, directly or indirectly, to and from the spacecraft surface or space. Because of the vacuum environment, conuectiue cooling normally used in electronic components is absent. Radiative transfer therefore plays a more important role than is the case in earth-bound electronics.
7.1
Fundamentals of Thermal Radiation The reader may already be familiar with m u c h of the material on the theory of radiative heat transfer that is presented in this section. Much more can be found in standard texts, such as in the book by Siegel and Howell [2]. We present some basic material here primarily to define the concepts and notation that will be needed later. Also, this section will provide the opportunity 269
C H A P T E R 7 Spacecraft Thermal Design
270
to discuss early some applications specific to space technology before describing the more elaborate m e t h o d s that are needed to calculate spacecraft temperatures. 7.1.1
Blackbody Thermal Absorption and Emission A blackbody is defined as an idealized material that absorbs all radiation incident on it, irrespective of the wavelength and of the angle of incidence. Reflection and transmission are therefore absent. As will be discussed here, the absorption and emission properties of a blackbody can be derived from one another, so that no new definition of the blackbody will be needed to describe its emission. The expression "blackbody" ~ which has historical r o o t s ~ i s somewhat inaccurate, because, as is clear from the definition, almost always only the surface of the body is important. Thus, if there is heat flowing by conduction to the surface, there will be a distribution of temperature in the body different from the surface temperature. Yet, conventionally, one still refers to the latter in this case as "blackbody temperature." There are no materials that m a t c h exactly the definition of a blackbody, although some ~ such as black paint in the visible range of wavelengths ~ come very close. Ideal blackbodies, however, are an important standard with which real surfaces can be compared. The following thought experiment, illustrated in Fig. 7.1, allows one to draw some simple conclusions from purely t h e r m o d y n a m i c arguments: One considers an evacuated enclosure at a uniform temperature and of arbitrary shape. The exterior is thermally insulated. The interior surface emits and absorbs as an ideal blackbody. Also, there is a test body (assumed for simplicity to be convex so that radiation from it strikes only the enclosure) which is also a black body. Initially, the temperatures of enclosure and test body are arranged to be the same. It then follows from the first and second laws of thermodynamics
Figure 7.1
Schematic illustrating derivation of Lambert's law.
7.1 Fundamentals of Thermal Radiation
271
that neither temperature can either increase or decrease and that the test body's emission equals its absorption. Because the argument is independent of the test body's position or orientation, the radiation in the space between enclosure and test body must be h o m o g e n e o u s (i.e., has constant radiation density) and isotropic. In Fig. 7.1a let dS be a surface element of the enclosure and fl the angle from the normal n of dS to some chosen direction of emitted radiation. Also let dw be the solid angle b o u n d i n g this direction. Whereas in Fig. 7.1a all rays emitted from a point of dS that fall within the solid angle dzo are shown, in Fig. 7.1b the rays from all points of dS are shown that converge to an aperture with area dB (located at point P at distance R.) Similarly, a second surface element dS' is shown, with rays converging to an aperture with the same area dB at P and at distance R'. The area dB can be chosen arbitrarily small c o m p a r e d with dS and dS'. Also, for convenience, dS and dS' are chosen such that the solid angles dfz extended from the apertures to dS and dS' are the same. The radiation from a point of dS or dS' striking the respective aperture therefore has directions b o u n d e d by the solid angles dw = d B / R 2
and
dw' = d B / R '2
(see Fig. 7.1.a). Another such geometrical relation, evident from Fig. 7.1.b, is d ~ = dS cos fl/R 2 = dS' cos fl'/R ~2
Let eb~ dS doJ dZ be the black body radiant power from dS, emitted within a cone with solid angle do centered about fi and within a wavelength interval d~. centered about Z. The eb~ is referred to as the directional, m o n o c h r o m a t i c emitted power (per unit area and unit wavelength interval; in the present case for a blackbody). It is a function of fl, ~., and T, and its magnitude is usually expressed in units of W / ( m 2 ~m). Analogous statements hold for eb~,~. Isotropy of the radiation at point P then requires that the two radiant powers be equal, hence eb~x dS d o dZ = eb~,~ dS' dw' d~.
Substitution of the geometrical relations found for dw, dw', and d ~ yields eb~ cos fl' = eb~,~ COSfl
(7.1)
In particular, w h e n expressing the power emitted in a direction given by fl with the power emitted along the normal, one obtains by setting fl' = 0 the important relation eb~z = ebnzCOSfl
(711 #)
The subscript notation used here and in what follows is more or less standard in the theory of thermal radiation. Thus n refers to the normal direction. Omitting a m o n g the subscripts the letter b m e a n s that a surface more general than a blackbody is considered. Omitting the angle fl or n m e a n s that the radiant power has b e e n integrated over all pertinent angles
272
C H A P -I- E R 7 Spacecraft Thermal Design (usually a half-space). Omitting the wavelength )~ m e a n s that the radiant power has b e e n integrated over all wavelengths. Equation (7.1') is known as L a m b e r t ' s cosine law (Lambert, astronomer a n d physicist, 1728-1777). The d e p e n d e n c e on fl o n l y t h r o u g h the cosine has a simple representation: As is easily shown, if the radiant power of a surface e l e m e n t is drawn as the lengths of vectors pointing along the directions of p r o p a g a t i o n of the radiation, the end points are located on a sphere that is t a n g e n t to the radiating surface. The radiant power, ebb, per unit area and unit wavelength interval, emitted into the half-space above dS, is k n o w n as the hemispherical, monochromatic emitted power (per unit area and unit wavelength interval) and is readily o b t a i n e d by integrating ebt~x over the half-space: ebx -- 27r ebnx
f
nl2 COSfl sin fl d~
rift=0
so that eb~ -- n~ ebn~
(7.2)
It was first shown by Planck (1858-1947), based on an electromagnetic calculation of the emission and treating the p h o t o n s in the enclosure as a perfect gas, that 2yrC1 e b ~ - )~5(exp(C2/~.T ) _ 1)
(7.3)
where C1 = hc 2 = 0.59544 10 -16 W m 2 and C2 = hc/k = 1.4388 104 # m K (h = Planck's constant = 6.6252 10 -34 J s; k = Boltzmann's constant = 1.3806 10 -23 kg m2/(s2IO; c = speed of light in v a c u u m = 2.99792 108 m/s). Comm o n units for eb~ are W/(m 2/zm), therefore the same as for eb~. Equation (7.3) is known as Planck's r a d i a t i o n law. Its derivation became the basis for the later d e v e l o p m e n t of q u a n t u m mechanics. In Fig. 7.2, the hemispherical m o n o c h r o m a t i c emitted power eb~ is plotted for several temperatures that are i m p o r t a n t in space technology. The curve labeled 5760 K corresponds to the effective blackbody temperature of the solar photosphere. Its radiation, to a large extent, determines the skin t e m p e r a t u r e of spacecraft. The curve labeled 300 K is representative of the majority of spacecraft c o m p o n e n t s . They are designed for, and work best, in a fairly narrow t e m p e r a t u r e range a r o u n d r o o m temperature (e.g., storage batteries in the range of 5 to 30~ The third curve applies to the boiling t e m p e r a t u r e (77 K) of nitrogen. This t e m p e r a t u r e is typical of m a n y scientific i n s t r u m e n t s that work at infrared wavelengths. Liquid nitrogen is also used in spacecraft thermal testing, where it cools the test c h a m b e r walls, thereby suppressing u n w a n t e d thermal radiation from the chamber. With increasing temperature, the m a x i m u m emitted power shifts toward shorter wavelengths. If)~max designates the wavelength for which at a chosen t e m p e r a t u r e the hemispherical m o n o c h r o m a t i c emitted power has a m a x i m u m , then )~max T = C3
(7.4)
where the c o n s t a n t C3 = 2.8978 103/zm K. First established experimentally,
7.1 108
273
Fundamentals of Thermal Radiation I
I
~
I
I
I
I
I
57
10 6
104 ebX W m 2 Bm
102 1 10 -2 10 -4 10 -6 I
0.01
[ 11
0.1
Ultraviolet
....
10
Visible
100 Infrared
1000 %(gm)
Figure 7.2 Planck's radiation law (Eq. 7.3) for three temperatures significant in space technology: (1) effective temperature of sun's photosphere, (2) ambient temperature, (3) liquid nitrogen temperature. but also an easily derived consequence of (7.3), the equation is known as Wien's d i s p l a c e m e n t law (Wien, 1864-1928). The factor eb~z of the directional, m o n o c h r o m a t i c emitted power, expressed by ebz from Planck's radiation law, becomes, by using (7.1) and (7.2), eb~;~--
~-1
eb~ cos fl
(7.5)
The h e m i s p h e r i c a l e m i t t e d p o w e r (per unit area), eb, is defined as the integral of ebz over all wavelengths. The integration over Planck's relation can be carried out in closed terms (although this is not immediately obvious). The result is the S t e f a n - B o l t z m a n n law, originally found by experiment, eb
---- a
T4
(7.6)
where = 2 C 1 ~ 5 / ( 1 5 C 4) = 5.6693 10 -8 W/(m 2 K4)
From this, the factor ebfi of the d i r e c t i o n a l e m i t t e d power, that is, the power emitted per unit solid angle in the direction defined by the angle fl, w h e n integrated over all wavelengths is ebf -- yr- ] COSfl r T 4
7.1.2
(7.7)
Solar Thermal Emission
The sun's thermal emission originates in the photosphere. Its temperature can be obtained from spectral m e a s u r e m e n t s and is found to be about 5950 K. Absorption by the layers above the photosphere modifies the radiation. Nevertheless, as illustrated in Fig. 7.3, the sun's thermal radiation into space is well approximated by blackbody radiation, albeit at the somewhat
274
C H A P T E R 7 Spacecraft Thermal Design 2400
,
,
2000
E ::k E
/ / /
1600
\
.
.
.
.
.
.
.
" \
'"/I
1200
.
9
\
'~
Sun Blackbody, 5760 ~K
-.
_
u cf~ Cl3 L_
800
4OO 0.2
I
0.4
I
0.6
I
0.8
I
1.0
I
1.2
I
1.4
I
1.6
I
1.8
2.0
Wavelength (gin)
Figure 7.3 Solar spectral irradiance at 1.0 AU compared with the blackbody effective temperature (5760 K) and with the blackbody temperature of the photosphere (5950 K). lower effective t e m p e r a t u r e of 5760 K. (Not shown in the figure are the Fraunhofer a b s o r p t i o n lines. Although s o m e of them, such as the p r o m i n e n t calcium K II line, are quite deep, they are far too narrow to show on the scale of the figure.) At the distance of 1 AU (the m e a n s u n - e a r t h distance), the sun's radius has an a p p a r e n t angle of 0.267 ~. Even at the position of Mercury, this angle is still only 0.69 ~. In applications to the t h e r m a l control of spacecraft, it can therefore be a s s u m e d that all rays of solar radiation that intercept a spacecraft c o m e from a point source and are in effect parallel to each other. At these distances, the sun a p p e a r s as a disk of uniform luminosity. This is consistent with Lambert's law: The geometric factor resulting from the projection of an e l e m e n t of the sun's surface along the line of sight, is just canceled by the cosine resulting from (7.1) or (7.7). It follows from conservation of energy that at distances large c o m p a r e d with the solar radius, where the sun can be a p p r o x i m a t e d by a point source, the intensity (i.e., the radiant power per unit area p e r p e n d i c u l a r to the ray), ]h, falls offas the inverse square ofthe distance. At earth vicinity, the intensity has b e e n m e a s u r e d by satellite i n s t r u m e n t a t i o n . It fluctuates s o m e w h a t during the course of a year, d e p e n d i n g on the earth's distance from the sun. Thus jh -- 1353 W/m 2 (the so-called solar constant) at 1 ALl = 1309 W/m 2 at aphelion (about July 4) = 1399 W/m 2 at perihelion (about January 3) Some of the solar radiation in the t h e r m a l range is a b s o r b e d in the earth's atmosphere, primarily by water vapor a n d carbon dioxide. However, this effect is negligible for t h e r m a l calculations pertaining to satellites, even w h e n the orbits are low (e.g., 300 km).
7.1 Fundamentals of Thermal Radiation
7.1.3
275
Thermal Emission, Absorption, and Reflection of Technically Important Surfaces The thermal radiation properties of technically important surfaces can differ greatly from those of the ideal blackbody surface. In principle, they could be obtained by q u a n t u m mechanical calculations. In practice, however, such calculations would be impossibly difficult to carry out, or else, if approximated, would be very inaccurate. Therefore such radiation properties can be obtained only by direct measurements. Whereas Eqs. (7.1) to (7.7) are thermodynamically precise, no such claims can be m a d e for the equations that follow (with the exception of Kirchhoff's law in its most general form). The thermal radiation properties of a surface are characterized by its emission, absorption, reflection, and transmission. Transmission hardly plays any role for surfaces that are employed in spacecraft and will therefore not be considered here. The transparent layers of fused silica or quartz that cover the surfaces of optical solar reflectors, or the covers of solar cells, are sufficiently thin that their temperature is virtually the same as that of their opaque substrate; for the purposes of thermal calculations they can therefore be considered as single, opaque units. Absorption and reflection d e p e n d not only on the properties of the surface but also on the direction and spectral characteristics of the incident radiation. For this reason, they are more difficult to take into account than emission. Even if a surface is isotropic, hence its emission i n d e p e n d e n t of the azimuthal angle, reflection does not need to be so because it will in general d e p e n d on the direction of the incident radiation. Opaque surfaces reflect what is not absorbed. The reflected c o m p o n e n t can often be approximated by the sum of a specular reflection and of a fully diffuse one. Let ir dS dzo d~ be the radiant power incident on dS, b o u n d e d within the solid angle dw of a cone that is centered on the incident direction, and for a wavelength interval d~. centered on the wave length)~. The angles r and fl are the azimuthal angle and the angle with the surface normal, respectively, of the incident radiation. For some purposes it is convenient to introduce in place of the incident power ir the intensity jr which is defined relative to a surface element with the same area, but perpendicular to the incident radiation, so that ir
- J~ex cos/3
(7.8)
Also, let correspondingly ac~x be the power per unit area, unit solid angle, and unit wavelength interval that is absorbed from radiation incident at the angles r and fl and wavelength )~. In all applications considered in this text, ar is proportional to ir (Nonlinear effects are important in the case of radiation from high-power lasers but are insignificant in the analysis ofradiation heat transfer between surfaces.) One then defines as the coefficient of proportionality the directional monochromatic absorptivity c~r by the relation acex = c~r
(7.9)
276
C H A P T E R 7 Spacecraft Thermal Design dS"
Black Isothermal Enclosure
Figure 7.4 Derivation of Kirchhoff's law. One also defines, relative to the blackbody emission ebbs, the directional monochromatic emissivity e ~ by the relation er
-- e,/~. eb/~x
(7.1 O)
For the remainder of this chapter it will be assumed that all surfaces, in addition to being opaque, emit and absorb isotropically and diffusely. (Fluorescent materials, i.e., materials that absorb at one wavelength and emit at a different, longer wavelength, hardly ever occur in engineering applications and are excluded in what follows.) The absorptivity and emissivity are related by Kirchhoff's law (Kirchhoff, 1824-1887), which can be derived by the following thought experiment: As illustrated in Fig. 7.4, a surface element dS of the sample material is placed at the center of a hemispherical enclosure ofradius R. With the exception of dS, all interior surfaces of the enclosure are ideal blackbody surfaces. The entire configuration is assumed to be initially at a c o m m o n temperature. The enclosure therefore is filled with a homogeneous, isotropic blackbody radiation. As is illustrated for the point P, but equally applies to all points of dS, the solid angle extended to a surface element dS' (an annulus on the hemisphere) is d~o = dS'/R 2. Similarly, the solid angle extended from a point P' of dS', or from any other point of dS', is do)' = dS cos fl/R 2, where fl is the angle of incidence of the radiation on dS from dS'. Therefore
dS' do)' = dS cos fi do)
(7.11)
The only radiation reaching dS within the solid angle do) is that coming from dS'. Hence the radiant power incident on dS within the intervals do) and d)~ is
i~ dS dw d)~ = ebn~ dS' do)' d)~ = ebn~ dS cos fl dw d)~ (keeping in mind the convention adopted for the notation of subscripts as introduced in Sect. 7.1.1). As before, it follows from the first and second laws of thermodynamics that the temperatures of dS and of the blackbody enclosure cannot change and that the absorption ofradiation by dS must equal its emission. Therefore a~x = e~x. Hence, with the definitions (7.9) and (7.10), ~
- s~
(opaque, isotropic surfaces)
(7.12)
277
7.1 F u n d a m e n t a l s o f Thermal Radiation
It follows from the definitions of the absorptivity and emissivity that for the ideal blackbody a~x - e~x = 1. It should be n o t e d that in the m o s t general case a~x a n d e~x are functions of the angle fl, the wavelength )~, a n d the t e m p e r a t u r e of the absorbing surface. There are, however, i m p o r t a n t instances in which s o m e or all of these functional d e p e n d e n c e s can be neglected. In particular, the d e p e n d e n c e on fl can be neglected w i t h o u t incurring a major error w h e n the surface is rough on the scale of the wavelength, as occurs with a majority of technical materials. Such surfaces reflect, absorb, and emit diffusively. It t h e n follows from (7.8) and (7.9) that (7.13)
a~z = a#z ]#z cos fi and from (7.10) and (7.1') that efik -~ 8fixebnx
(7.14)
COS fl
showing that the absorption a n d emission of an ideal diffuse surface have the same cos fl d e p e n d e n c e as is the case in Lambert's law (7.1') for blackbody emission. (Actual diffuse surfaces show deviations from the ideal cosine d e p e n d e n c e near the glancing angle, i.e., near fl = 90 ~ For metals a~z and e~z are s o m e w h a t larger there, for insulators s o m e w h a t smaller t h a n what would be indicated by the simple cosine dependence.) In w h a t follows, diffuse surfaces will be a s s u m e d t h r o u g h o u t this chapter. Another form of Kirchhoff's law is o b t a i n e d by integrating the angle fi over the half-space above the diffuse surface. In a n o t a t i o n that is selfexplanatory, one defines a ~ a n d ez by ax = ax i x
(7.15)
ex =
(7.16)
sxebx
The coefficients az and ez that are defined by these equations are referred to as the m o n o c h r o m a t i c a b s o r p t i v i t y and e m i s s i v i t y , respectively. Hence, because a~z and e~z are n o w i n d e p e n d e n t of fl, ax -- a#x L
.S.
i#x do) - a#x ix,
ex -- e~x L
.S.
eb~x do) -- e#xebx
(h.s. = half-space) so that a#z = az a n d e ~ = ez. Therefore, from (7.12), az = ex
(opaque, isotropic, diffuse surfaces)
(7.17)
A third form of Kirchhoff's law is o b t a i n e d by integrating (7.15) a n d (7.16) over the wavelength. One t h e n defines the absorptivitya a n d emissivity e by a = aL
e = eeb
so that from a -
i
X2
a~ ix d)~,
1
e - fA A2 e~eo~ dk i
(7.18)
C H A P T E R 7 SpacecraftThermal Design
278
follows c~ --
f~a azi~ d~.
e-
f~l2 ix d&
fAa2 exeb~ d~.
(7.19)
faA2 eb~ d~.
It should be noticed that in spite of Kirchhoff's law in the form (7.17), it is in general not true that c~ = e. The reason is that the weighing factors used in defining c~ and e d e p e n d on the spectra of the incident, respectively emitted, radiation. As is evident from the examples shown in Figs. 7.2 and 7.3, the weighting factors ix and ebx a p p r o a c h zero as )~ ~ 0 and c~. To c o m p u t e ct and e from (7.19), it is therefore sufficient to know a~ in a wavelength interval ~,1 to ~,2 and e~ in the interval A ~to A2, that is, in an interval where most ofthe radiant power is concentrated. For instance, for the blackbody emission, it is easily established by Planck's radiation law (7.3) that the emitted power below XT = 1450 lzm K and above 23 200 # m K is only 1% of the total. In the case of solar radiation, with T = 5760 K, the corresponding limits of integration are 0.25 lzm for the lower and 4.0 tzm for the upper limit. Assuming 300 K for the temperature of a spacecraft c o m p o n e n t , these limits are 4.8 and 77 # m . It is i m p o r t a n t to note that for the absorptivity the limits d e p e n d on the temperatures of the sources of the radiation. Surfaces for which there is a common interval of wavelength in which m o s t of the radiant power is found and in which a~ and e~ are approximately independent of the wavelength are called (somewhat misleadingly) gray surfaces. In that case, it follows from (7.17) that for a given material at a given t e m p e r a t u r e a = e
(opaque, isotropic, diffuse, gray surfaces)
(7.20)
In the analysis of radiant heat transfer b e t w e e n spacecraft components, the t e m p e r a t u r e s are m o s t often in a relatively narrow range, say from 270 to 350 K. The limits ofintegration for ~ and e in (7.19) then can be taken to be the same. M1 other conditions for the validity of (7.20) are also a s s u m e d satisfied. In what follows, this temperature range will be referred to as ambient and designated by the subscript ( )a. Of course, for solar radiation, designated by ()h, the wavelength interval in which m o s t of the power occurs is quite different from that for absorption by surfaces that are at a m b i e n t temperature. To summarize: 0Ca "~ ea
but
0/h ~ O/a
(7.21)
[Of course, eh w o u l d be only of astrophysical interest; for spacecraft applications, instead, it is only the solar constant (Sect. 7.1.2) that matters.]
7.2
Spacecraft Surface Materials In Fig. 7.5, a d a p t e d from Agrawal [4], Ofa = 8a (the absorption and emission coefficients at r o o m temperature) as functions of the wavelength, are shown for three different surfaces that are frequently used in applications to
279
7.2 Spacecraft Surface Materials 1.0 0.8 - i ' i cO
0.6
,
i
i'
:
S13-G-LO
i '--White Paint
|
,
E
-I'
0.4 0.2
Paint
|
,---'
_~'
"'-' ," '"
i
:
i
j
./'i/'. !
1
|
ii
3 Wavelength (gin)
,rror
_
V-
i
/X_ Second
I
0.3
|
!/"
Surface
I
10
30
Figure 7.5 Monochromatic absorptivity/emissivity at ambient temperature at beginning of life for three typical spacecraft surfaces. From Ref. 4, Agrawal, B. N., "Design of Geosynchronous Spacecraft," Prentice-Hall. Copyright 9 1986 Prentice Hall, Inc. By permission. space technology. White and black paints are used on external surfaces of spacecraft. Electronic boxes are usually painted black to promote radiative transfer among the various components interior to the spacecraft, thereby equalizing to the extent possible their temperatures. Optical solar reflectors (OSRs) are second surface mirrors, often made from a thin sheet of fused silica, silvered on the back. They strongly reflect solar radiation, yet emit well in the infrared. They are therefore used to cool by radiation into space components that have high rates of electric-thermal dissipation, such as is the case for traveling wave tubes. Notable are the strong absorption and emission bands in the infrared, resulting from the vibrational bands of surface atoms. Paint that is normally white is essentially black at wavelengths above 3 lzm. That the difference between C~a -- Sa and ~h can be very large is seen for white paints. When they are first exposed to the space environment, the absorptivity C~hm a y v a r y from 0.2 to 0.3, whereas aa may vary from 0.8 to 0.9. Figure 7.6, adapted from Wingate in Pisacane and Moore [5], and from Schmidt in Hallmann and Ley [6], indicates the range of aa = ea and C~h of various materials and coatings that are used in spacecraft design. (As before, the data apply to surfaces when first exposed to the space environment.) It is noteworthy that all four corners of the diagram, that is, all extreme combinations of the coefficients, can be approximately realized by the proper choice of materials. Intermediate values of the coefficients can be obtained with other materials or, as is frequently done, by applying to the spacecraft surface aluminized or other tapes in patterns of alternating absorptivity that then produce the desired mean absorptivity. The application of tapes to surfaces of spacecraft is also a convenient method when there is a need to correct discrepancies between design temperatures and the true temperatures measured in thermal vacuum tests. Although C~aand ea change relatively little with time of exposure to the space environment, this is not the case with solar absorptivity. Particularly
280
C H A PT E R 7
SpacecraftThermal Design 1.0
I
Black Ni, Cr
[~
0.8
Mg Anodized Black --~ Solar Cells ~
[] j [ ]
Black Paint
[~
r
06
Be Polished ~ ] Aust. Steel Polished
O_ 0 ./D
<
0.4 -
0
0.2
Au
I
Tiodized Ti O
AI Paint ~~_~z-~ L . ~ ~
Aluminized Kapton Ik
Wahiitn~
( ~ Mg Anodized Anodized AI~ I~-'-"-'-i FEP Teflon with Ag, AI
0
-
~ I
0.2
I
I
I
04 0.6 0.8 Emissivity ca = % Absorptivity
1.0
Figure 7.6 Solar absorptivity versus ambient temperature absorptivity/emissivity at beginning of life. OSR = optical solar reflectors. After Wingate [5] and Schmidt [6]. Adapted from and added to from Ref. 5, C. A. Wingate in "Fundamentals of Space Systems," Pisacane, V. L. and Moore R. C. Copyright 9 1994 by Oxford University Press, Inc. Used bypermission of Oxford University Press. surfaces that initially have a low value of ah, such as white paints and to a lesser extent anodized aluminum, suffer major increases in solar absorptivity over time (e.g., over the 5 to 10 years of the useful life of m a n y spacecraft). The most important contributor to the deterioration of the surface is usually the sun's ultraviolet radiation. Micrometeoroids, in the long term, also affect the surface properties by causing dings in the surface. In the near-earth environment, typically even more important is small space debris, particularly aluminum oxide particles from solid-propellant motors that had been fired in the past at the vehicle's altitude. Although effective micrometeoroid shields have been developed for use on space stations, particularly on the sides receiving the largest n u m b e r of impacts, the shields themselves are also subject to the long-term increase in solar absorptivity. Therefore there may still be a requirement for active thermal control. For optical solar reflectors, Teflon-based second-surface mirrors have been found to be particularly useful. But even they show some significant increases of solar absorptivity when the dose of 5 keV to 1 MeV protons and electrons exceeds about 101S particles per cm 2. In the case of low-altitude earth-orbiting spacecraft, it has been found that coated Kapton and Mylar deteriorate quickly as a consequence of the chemical interaction of the material with the atomic oxygen in the upper atmosphere.
7.3
Model of a Spacecraft as an Isothermal Sphere
281
Still another cause of increased solar absorptivity over time is the effect produced by outgassing of organic spacecraft materials, such as electrical insulators. When vented to the outside, the vapors can be adsorbed on exterior surfaces of the spacecraft, thereby increasing the absorption of solar radiation. A frequently employed remedy is intentional venting of the vapor at an opening in the spacecraft skin where vapor adsorption next to the vent is of less concern. The result of these effects is a tendency toward an increased temperature of the spacecraft. The designer of long-life spacecraft must take account of this by planning for a relatively low temperature at the beginning of the mission or by providing active thermal control (discussed at the end of this chapter). Toward the end oflife ofthe spacecraft, its operator may also have to reduce the operating level to minimize the thermal dissipation of one or more of the spacecraft's electronic and communications components.
7.3
Model of a Spacecraft as an Isothermal Sphere We consider a spherical shell of radius R, uniform temperature T, uniform solar radiation absorptivity ah, and uniform emissivity ea. Electric-thermal dissipation in the interior of the sphere is (at first) neglected. Except for several spherical shells that have been orbited for the purpose of radar calibration, spacecraft of course are not spherical but come in a great variety of shapes. Considerations of isothermal spheres, however, provide a simple means for a comparison of the effect that different surface coatings can have on the overall temperature of vehicles. By integrating (7.13) over a hemisphere, or more simply by noting that the incident radiant power is proportional to the circular disk formed by the (parallel) incident solar rays, the absorbed power is Jr R2othjh. In the case of thermal equilibrium, the absorbed power is equal to the emitted power 4Jr R 2 eaO" T 4. Solving for T gives
T_ (c~hjh) 1/4 4eaO"
(7.22)
The equilibrium temperature therefore depends only on the ratio of Oth and ea. (It follows from equating absorbed and emitted radiant power for an isothermal body of general shape in a fixed orientation to the sun that also in this case the equilibrium temperature depends only on the same ratio.) Table 7.1 shows some equilibrium temperatures for spheres at average distances from the sun for Earth (1 AU), Venus (0.723 AU), and Mars (1.523 AU). Next we consider the temperature change of the sphere as it b e c o m e s eclipsed and as it exits again from the eclipse. Such is the case with geostationary satellites. These spacecraft are periodically eclipsed each spring and fall at and around the equinox. The eclipse lasts a m a x i m u m of 72 minutes [taken from the center of the first (entering) p e n u m b r a to the center of the second (exiting) penumbra]. This m a x i m u m duration occurs at equinox.
282
C H A PT E R 7
Spacecraft T h e r m a l Design
Table 7.1 Equilibrium Temperatures of Spheres Equilibrium temperature (K) of spheres at
White paint, BOLa EOL Black paint Gold
C~h
Ca
Cgh/ Ca
Earth
Venus
Mars
0.2 0.6 0.9 0.08
0.9 0.9 0.9 0.03
0.22 0.67 1.00 2.67
191 251 278 355
365 480 532 679
82 108 120 153
a BOL, Beginning of life; EOL, end of life.
As before, we consider an isothermal, thin shell of uniform emissivity and absorptivity. Electric power Pel is a s s u m e d to be generated with conversion efficiency/]el by a solar panel equal in size to the projected area of the sphere. Therefore Pel = rr R2~eljh. The generated power is a s s u m e d to be dissipated in the interior of the vehicle, b o t h out of and during the eclipse, in the latter case supplied by storage batteries. If m is the mass of the shell and c the specific heat per unit mass of the material, steady-state energy balance requires that Jr R2C~hjh -- 47r R2ea a T 4 -t- Pel
mc dT/dt -
--- 7~R 2 [(o~h -~-/]e)jh -- 4eaa T 4]
(out of eclipse)
- 4 7 r R 28ao" T 4 + Pel
mc dT/dt -
(in eclipse)
= 7r R 2 [/]e jh -- 4 e a a T 4]
(7.23a)
(7.23b)
If Th and T-h designate the temperatures out of and in the eclipse, as they would be if steady-state equilibrium prevailed, then 'j4
-
4~aO"
T-h --
'
[e'l'J4 48aO"
(7.24)
Also, let rh and r - h designate time constants defined by me
•h --
47r R28a a
Z'-h --
T~'
me
47r R28a O"T3h
(7.25)
Using these quantities, the energy balance equations b e c o m e dt
-
Z'h
dT dt
=
"v_h
T-h
Th
E
1 -
T ~
4
(out of eclipse)
(7.26a)
(in eclipse)
(7.26b)
therefore two nonlinear equations. Their solutions are coupled by the req u i r e m e n t s of continuity of the t e m p e r a t u r e at the entrance [designated by the subscript ( )1] and exit [designated by ( )2] from the eclipse.
7.3 Model of a Spacecraft as an Isothermal Sphere
283
Then from
f
dT 1 - (T/Th) 4 rh (tanh- 1 T~ Th + t a n - 1 T~ Th) + const., 2
f
T < Th
dT 1 - (T/r_h) 4 T-h 2
(coth -1 T~ r-h
-
-
cot -1 T~ T - h )
-[-
const.,
T>
T_ h
follows, observing that (7.26a) and (7.26b) are each separable, that t a n h -1 T1/Th + tan -1 T1/Th -- tanh -1 T2/Th -- tan -1 T2/Th = 2[p-- (t2 -- tl)l/rh c o t h -1 T2/r-h
--
cot -1 T2/r-h
= 2[t2- tl]/r-h
(7.27a) --
coth -1 r l / r - h
-[-
COt-1 rl/Z-h (7.27b)
where a periodic dipping, with period iv, in and out of the eclipse has been assumed. T1 and T2 represent the extremes in the cyclic temperature variation. They can be found by solving the two coupled, transcendental equations (7.27) by Newton's or one of the related numerical methods. To provide a general estimate of the magnitude of the temperature fluctuation, the following example is useful: The sphere is assumed to be in geostationary orbit at the time of equinox, so that t2 - tl = 72 minutes, p = 24 hours. The shell is assumed to be 3.4 m m thick, with a specific heat per unit mass of 960 J/(kg K) and density 2800 k g / m 3 (for a l u m i n u m 2014-T6), solar radiation absorptivity 0.55, ambient temperature emissivity 0.30, and an electric conversion efficiency of 0.15. The extremes of the temperatures, as calculated from (7.27), are T1 = 343 K (virtually the same in this case as the steady-state equilibrium temperature Th) and T2 = 290 K, resulting in a m a x i m u m temperature swing of 53 K. The result is i n d e p e n d e n t of the radius of the sphere. In this calculation, instantaneous transitions of the spacecraft from being in the sun to being totally eclipsed, and back again, were assumed. In fact, as shown schematically in Fig. 7.7, before entering the u m b r a (Latin, shadow), where it is completely eclipsed, the spacecraft passes briefly through the penumbra, where the earth blocks the solar disk only partially. The crossing by a geostationary satellite ofthe p e n u m b r a takes 2.1 minutes at equinox. After being completely eclipsed, the spacecraft passes the p e n u m bra once more. For the purpose of calculating the electrical power available from solar panels and for calculating the spacecraft temperature, the modifications to the assumed instantaneous transition in and out of an eclipse are negligible.
C H A P T E R 7 Spacecraft Thermal Design
284
Figure 7.7 Schematic of diurnal and annual rotations of the earth and of a threeaxis stabilized, geostationary spacecraft, shown at local midnight. S, sun; Ec, ecliptic; C.Eq.P., celestial equatorial plane, V.E. (A.E.), vernal (autumnal) equinox; S.S. (W.S.), summer (winter) solstice; E.P., W.P., N.P., east, west, north panel. 7.4
Earth Thermal Radiation and Albedo The earth's surface, atmosphere, and clouds radiate into space at infrared wavelength (as corresponds to their relatively low temperature). At its surface, the earth is approximately in thermal equilibrium between the power absorbed from the incident solar radiation and its thermal emission. By comparison, the flow from the earth's interior of heat, p r o d u c e d by the radioactive decay o f u r a n i u m and its daughter products, is only a m i n o r contributor. The earth's t h e r m a l r a d i a t i o n follows roughly Lambert's law. The emitted power per unit area (outside the a t m o s p h e r e and averaged over the earth's surface and time) will be designated by eg. It has been m e a s u r e d by spacecraft i n s t r u m e n t a t i o n at 237 + 7 W / m 2. For low-altitude satellites, the earth's thermal radiation can be an imp o r t a n t contributor to the spacecraft's temperature. Taking as an example a satellite at an altitude h = 300 km, it is easily calculated from the geometry of the satellite position that the spacecraft has a field of view of the earth that extends in arc length from the nadir to a b o u t 6.4 times the altitude. For the p u r p o s e of heat transfer calculations, the earth's surface can therefore be replaced approximately by an infinite plane. To obtain a rough estimate of the effect, we represent the satellite as a sphere of radius R a n d absorptivity OCafor a m b i e n t temperature incident radiation. If dS = r de dr (r = distance from the nadir, r = azimuthal angle) is a surface element of this plane, then, since the d i m e n s i o n s of the spacecraft are negligible c o m p a r e d with its altitude, the solid angle dw extended from dS to the spacecraft is Jr R2/(h 2 4. r2). The power emitted by dS, incident on the half of the sphere that faces dS, is eg dS cos fl Jr R 2/ (h2 4- r 2) where fl is the angle from the local vertical to the line from dS to the satellite. Therefore, with 0 = r~ h, the total power, Ag, a b s o r b e d by the satellite from earth thermal radiation is Ag -- yr R2c~aeg h
r dr de
Q do de
fr~=0 ~ =0 27r (h2 4- r2) 3/2 = 7/"R2c~aeg f0~=0 ~ 27r =0 (1 4- 02) 3/2 -- 2yr2R2c~aeg
(7.28)
7.5 Diurnal and Annual Variations of Solar Heating
285
Comparing this result with the absorption, Ah, from direct solar radiation (Sect. 7.1.2), the ratio is seen to be Ag
27r. 237 (W/m 2) O~a
Ah
1353(W/m 2) O~h
Cga ~ 1.10 m O~h
(7.29)
For low-altitude satellites, depending on the ratio Ola/Olh, the earth thermal and direct solar radiation therefore can be of comparable magnitude. At higher altitudes, earth thermal radiation becomes increasingly unimportant. For instance, in the case of a geostationary satellite (orbit radius = 42,164 km) the earth can be approximated roughly as a point source. A simple calculation then shows that the factor of 1.10 in (7.29) is replaced by 0.004. For geostationary spacecraft, earth thermal radiation therefore is negligible. The method for calculating the effect of earth thermal radiation on spacecraft of arbitrary shape is virtually the same as the method outlined in Sect. 7.8 for solar radiation. Calculations analogous to those discussed for the earth apply to other planets and to the m o o n as well. In the case of the outer planets, their surface temperatures are so low that the effect of their thermal emission on a spacecraft becomes negligible. The solar radiation incident on the earth is in part absorbed and in part reflected. The latter is referred to as e a r t h albedo (Latin albus: white). The maximum effect on the spacecraft occurs when sun, spacecraft, and earth are approximately aligned. The albedo is a highly variable quantity because it largely depends on the presence or absence of clouds. At optical wavelengths, clouds, snow, the polar ice, and the oceans all have high reflectivity compared with land masses. Reflected radiation is frequently approximated by the sum of a specular and a diffuse component, but predictions of the albedo are difficult because, in addition to depending on meteorological data, they depend on the position of the spacecraft relative to the sun. The reflected radiant power per unit area can be expressed by the product agjh where jh -- 1353 W/m 2 is the solar constant (Sect. 7.1.2) and ag is known as the earth albedo coefficient. As an average over the satellite orbit, the annual m e a n value of ag is about 0.3, somewhat lower for equatorial orbits and somewhat higher for polar orbits. Like the earth thermal radiation, and for the same reasons, albedo radiation is important only in the case of low-altitude spacecraft. If the satellite always faces the earth with the same side, as is the case with three-axis stabilized vehicles, the effect on the spacecraft of the variability of the albedo can be reduced by selecting materials for the earth-facing side that have low solar absorption coefficients.
7.5
Diurnal and Annual Variations of Solar Heating The thermal input to a spacecraft by solar radiation over the course of 24 hours and over a year is illustrated in Fig. 7.7. The figure applies to a geostationary, three-axis stabilized vehicle. Seen from the north, the earth's daily rotation about its axis is anticlockwise. So is the direction of the earth's annual orbit about the sun in the
C H A P T E F~ 7 SpacecraftThermal Design
286
ecliptic plane. Geostationary satellites, as the name implies, orbit the earth in the equatorial plane synchronously with the earth. The equatorial and ecliptic planes include an angle of 23.5 ~ referred to as the obliquity of the ecliptic. The four positions of the satellite shown in the figure are at local midnight, that is, w h e n the spacecraft's projection onto the earth is at 24.00 hours local solar time. At the time of the vernal or autumnal equinox, that is, when the earth crosses the line of intersection of the ecliptic and celestial equatorial planes, the sun, earth, and spacecraft in its midnight position are collinear, and the spacecraft is eclipsed. At s u m m e r or winter solstice, there will be no eclipse because the distance of a geostationary spacecraft from the earth is sufficiently large for the spacecraft to pass the earth's u m b r a and p e n u m b r a below or above the ecliptic plane. However, near the equinox positions there will be about 90 days per year, centered about the equinox and the spacecraft midnight positions, where the spacecraft will be eclipsed. The longest duration of the eclipse is about 72 minutes. Three-axis stabilized satellites rotate about one of their axes so as to point their high-gain antennas at all times toward the earth. For the solar panels to point toward the sun, they must rotate relative to the body of the spacecraft through 360 ~ every 24 h. In the usual arrangement, the axis of rotation of the solar array is normal to the equatorial plane, so that the angle of incidence of the solar radiation stays roughly constant at 90 ~. By a nomenclature that has become conventional for three-axis stabilized satellites, the sides of the main body are called east, west, north, south, earth, and antiearth (or space-facing) panels. The north and south panels, because solar radiation is incident on them at most at a glancing angle of nominally 23.5 ~ have the advantage over the other panels of having a nearly constant, cool temperature. They are therefore favored for the placement of electronic components that have large heat rejection.
7.6
Thermal Blankets Thermal blankets, also referred to as multilayer insulation, are used on spacecraft to insulate thermally against radiative heat transfer. They are used in such applications as insulation against solar heating and against heat transfer from hot motors and exhaust plumes. They can also serve as a thermal barrier during the launch phase when the spacecraft is no longer protected by the launch vehicle's payload shroud. Thermal blankets are used not only to reduce heat gain but also to reduce heat loss. Particularly in deep-space missions, electric heating of key c o m p o n e n t s may be needed. The application of multilayer insulation then reduces the need for electric power. As shown in Fig. 7.8, thermal blankets consist of a n u m b e r oflayers, most often aluminized Mylar or Kapton, typically 20 to 30 # m thick. For higher temperatures, up to 340~ Kapton is preferred over the less costly Mylar. For the outermost layer of blankets that must insulate against radiation from hot rocket motors, titanium or stainless steel thin sheets can be used. Such sheets can withstand for short periods temperatures of up to 1400 ~C.
7.6 Thermal Blankets
287
k=l 2 2
H/1/1/1/I///////////////////
Z_ Low Conductivity Mesh
Figure 7.8
Thermal blanket for insulation against solar radiation. Ideally, the outer surface of a blanket that insulates against radiant heat input should have an absorptivity as low and an emissivity as high as possible. A compromise, however, is needed where high-temperature-resistant materials must be used. To minimize thermal conduction from layer to layer, a spacer is used to prevent contact of the sheets. As shown in the figure, this m a y take the form of a mesh of low-conductivity material, such as Dacron or glass fibers. Crimping or embossing ofthe sheets can serve the same purpose by reducing the contact areas. Because of the difficulty in defining the conduction path with any accuracy, the part played by conduction through the m e s h must be determined experimentally. An additional consideration in designing thermal blankets is that perforations of the sheets are needed to vent, during launch, the air otherwise trapped between layers. Because the thermal capacity of the layers is very low, a quasi-steady calculation that assumes a steady-state temperature distribution at each instant of time gives sufficiently accurate results. Also, because the separation between layers is very small c o m p a r e d with their lateral dimensions, edge and curvature effects of the blanket can be neglected. Except for the outermost surface, which usually has a different coating, we assume the same absorptivity O/a (-- 6a) for all i n w a r d - f a c i n g surfaces and the absorptivity a a for all o u t w a r d - f a c i n g surfaces of each layer. The temperature of the kth layer is designated by Tk. Let i0 be the radiant power incident on the thermal blanket. The solar absorptivity of the outermost surface will be designated by C~h0,itS emissivity by Sa0. Also let ik, k+l be the radiant power, s u m m e d over all directions and wavelengths, that crosses in the inward direction a control surface between layers k and k § 1. It therefore includes the emission from the inner surface of layer k, as well as all reflections from it. Similarly, ik+l,k is defined as the radiant power crossing the same control surface in the outward direction. Therefore ik, k+l -- 8aO" T ; -[- (1 -Ola)ik+l, k = a a a T/~ nt-
(1 - a a ) i k + l , k
(k = 1, 2, . . . . n -
1)
288
C H A P T E R 7 Spacecraft T h e r m a l Design where the last term is the radiant power reflected from the outward-facing surface of the (k + 1)th layer. Similarly, ik+l,k
-
8taO"r ; + l - [ - ( 1 -
Ol~a)ik,k+l
!
---- OCaO"T~+ 1 A- (1 - Olfa)ik, k+l
( k = 1, 2 , . . . , n - 1)
Eliminating ik, k+l and ik+l,k and taking the difference q = ik, k+l -- ik+l,k gives for the net inward thermal flux (7(T~-
q -
Tk41)
1/0r a -[- 1 / a ' a - 1
(k = 1, 2 . . . . .
n-
1)
(7.30)
which, by conservation of energy, is the same for all spaces between the layers and equals the net thermal power per unit area that reaches the component wrapped in the blanket. Adding these equations results in
T4) (n- 1)q-
(7.31)
1/C~a + 1/oe a -- 1
Using the thermal balance for the outermost layer q = o~hoio -- 8aOa T14
(7.32)
[1 + ( n - 1)eao(1/C~a -+- 1 / a ' a - 1 ) ] q - C~h0i0- 8a0O"T2
(7.33)
to eliminate T~ results in
where the right side (7.34)
qO -- CghOiO -- 8aO O" T ;
is recognized as the heat transferred to the c o m p o n e n t if no blanket was used and if the c o m p o n e n t surface had the absorptivity and emissivity of the outermost surface. A c o m m o n l y encountered problem is one in which Tn is prescribed and one wishes to calculate q and T1 to Tn-1. Thus, Tn might be the boiling temperature of a cryogen or the specified m a x i m u m allowed temperature of the component. The final result is therefore conveniently written as __q =
q0
1 1 + ( n - 1)ea0(1/aa -F 1 / % -
1)
(7.35)
It follows that for an effective insulation the solar absorptivity of the outermost layer should be as low as possible, and the emissivity at ambient temperature as high as possible, compatible with material limitations. It also follows that there is no theoretical advantage in choosing aa and a a differently; both should be as small as possible. The temperatures of the individual layers are readily obtained from (7.30) by finding Tn-1 from the last equation, then Tn-2 from the second to the last, and so on.
7.7 Thermal Conduction
7.7
289
Thermal Conduction Heat transfer a m o n g spacecraft c o m p o n e n t s is in part by radiation and in part by conduction. (Convective heat transfer, as it applies to rocket motors, is discussed in Sect. 4.14.) The reader is likely to be already familiar with the fundamentals of thermal conduction. Much material is readily available in several standard texts, such as the one by Carslaw and Jaeger [7]. The present section therefore merely lists some basic relations and adds some c o m m e n t s that apply to spacecraft. For an isotropic material, the rate of heat, q, conducted per unit area is q - -kgrad T
(7.36)
where k is the t h e r m a l conductivity. In some applications, particularly w h e n cryogens are part of a system, it is necessary to take into account the d e p e n d e n c e of k on the temperature. Thus, elemental a l u m i n u m has a thermal conductivity of 210 W/(m IO at 300 K, but 420 W / ( m K) at the normal boiling temperature (78 K) of nitrogen and 5700 W / ( m K) at 20 K. This last temperature is typical for m a n y spaceborne scientific instruments operating in the far infrared. Including time d e p e n d e n c e and heat production, conservation of energy requires that divq = -Oc ~T/~t + Ow
(7.37)
where 0 is the density of the material, c its specific heat per unit mass, and w the rate of internal heat produced per unit mass. Examples of the latter are the heat production that occurs in electrical conductors and in radioisotope sources used in deep-space probes for energy generation. It follows from (7.36) and (7.37) that 1
V2T + ~ grad k. grad T
10T a20t
t- kW - 0
(7.38)
where a 2 - k~ (Oc), called the t h e r m a l diffusivity. It is a measure of the rapidity with which temperature changes propagate through the material. Fourier's heat conduction equation A 2
T
.
10T . . a 2 0t
.
0
(7.38')
is an important special case, obtained w h e n there are no internal sources of heat generation and w h e n k can be assumed to be constant. A substantial n u m b e r of exact solutions of (7.38'), satisfying various b o u n d a r y conditions, are known. In spacecraft design, they have been largely replaced by numerical methods that are based on finite-difference or finite-element methods. Software is available that, once the geometry and b o u n d a r y conditions have been specified, allows one to obtain rapidly numerical solutions that would be unattainable by classical analysis. The same software also provides for presenting the results graphically for inspection by the spacecraft designer.
C H A P T E R 7 Spacecraft Thermal Design
290
Traveling Wave Tube Amplifier
Collector Housing
Thermal Doubler Honeycomb Skin
Optical Solar Reflectors
Figure 7.9 Heat rejection from a traveling wave tube amplifier by means of a thermal doubler and optical solar reflectors. The functioning of integrated circuits and of their assembly into circuit boards depends critically on the removal of heat by conduction. Because on spacecraft the electronic enclosures are usually vented, convection is absent. Typically, the enclosures are attached to a l u m i n u m shelves whose function is not only to support the units during launch but also to remove by conduction the heat generated by electric-thermal dissipation. To remove the heat from devices with high power consumption, for instance traveling wave tube amplifiers, t h e r m a l doublers are used to conduct the heat from the relatively small device to a larger area that then radiates the heat m directly or indirectlym into space. An example is shown in Fig. 7.9. In this case the principal heat source is the anode or collector of a traveling wave tube. The conduction path leads from the collector and its housing to an a l u m i n u m alloy thermal doubler and from there through the spacecraft skin to optical solar reflectors. These are typically small, about 5 by 5 cm silica glass squares, at most a few millimeters thick, silvered on the back side, and attached to the spacecraft skin by adhesive. Joints between two metal surfaces that are bolted together sometimes present an unwanted barrier to heat transfer because actual metal-to-metal contact may be confined to small areas only, often just around the bolt holes. In such cases, to improve the heat transfer, special greases with high thermal conductivity and low vapor pressure can be used.
7.8
Lumped Parameter Model of a Spacecraft The design of a new spacecraft requires extensive thermal calculations. Their purpose is to make certain that the specified temperature limitations of the various c o m p o n e n t s are not exceeded at any time during the life of the spacecraft. Such calculations must be performed for the different orientations of the spacecraft relative to the sun as they may occur during normal operating conditions. The same types of calculations are also carried out for times w h e n the spacecraft is eclipsed. During the launch phase, after the launch vehicle's shroud is removed and the spacecraft released, solar array panels, antennas, and booms are often still folded against the main body. The thermal configuration then m a y be quite different from the normal operating condition with its deployed panels and antennas. Separate calculations are also needed for this case.
7.8 Lumped Parameter Model of a Spacecraft
291
The thermal calculations go h a n d in h a n d with the design process. Their results will indicate required design changes. Conversely, design changes will induce modifications of the thermal calculations. Later in the buildup of the vehicle, components, subsystems, and the entire spacecraft (in the case of large craft often without solar panels, for reasons of size limitations of existing test facilities) will undergo thermal tests. For thermally sensitive components, discrepancies between test and calculation of no more than about 10~ are usually acceptable. Larger discrepancies require an investigation of their cause and corrective action. The latter may be in form of a modification of the design or, if there is separate justification for it, by changes in the thermal model.
7.8.1
Basic Relations
Thermal models, such as the simplified version illustrated in Fig. 7.10, divide the spacecraft into nodes, each assumed to have uniform temperature, absorptivity, and emissivity. The nodes interact with each other thermally by radiation and conduction. To obtain the required accuracy, depending on the complexity of the spacecraft, one hundred or more such nodes are often needed. To some extent the choice of nodes is arbitrary; comparison with test results and
I
:
//-42
~tion
~1
14 12
_
L ~
/~ / /
ntennas I
/~
I
"/
ApogeeMotor
ThrustCone ....
PropellantTank
17-40
2 ISolarPanel
Figure 7.10
Model for the thermal analysis of a spacecraft.
292
C H A PT E R 7
Spacecraft Thermal Design
resulting modifications of the model will improve the choice. Thermally sensitive subsystems will require a more extensive breakdown into nodes than is needed for those less sensitive. Typically selected as nodes are the main structural elements of the spacecraft, electronic boxes, solar array panels, antennas, rocket m o t o r components, temperature-sensitive components such as storage batteries, propellant tanks, cryogenic subsystems, and scientific instruments. Still others are thermal control devices such as louvers and those using phase-changing materials. As formulated in the following, the nodes are defined such that the radiative exchanges are between one-sided surfaces, such as the outer surface of an electronic enclosure. Panels, such as shelves or solar arrays, may have different absorptivities and emissivities on the two sides but typically are thin enough so that the temperatures of either side are virtually the same. They are represented by two different nodes, although at the same temperature. In what follows, all surfaces are assumed to be diffuse and their emission, absorption, and reflection to obey the simple cosine dependence on the angle between the direction of the radiation and the normal, in accordance with (7.13) and (7.14). (The active sides of solar panels reflect the incident solar radiation at least in part specularly. But because of their orientation toward the sun, this radiation does not reach other parts of the spacecraft, hence has no effect on the temperature of the latter.) Solar radiation is included in the following formalism. For simplicity, earth thermal and albedo radiations are omitted. However, if needed, earth thermal radiation can be dealt with by introducing the earth as just one more node in addition to the spacecraft nodes. Albedo radiation, because its spectrum is similar to the solar spectrum, can be treated in the same way, as will be carried out here for the incident solar radiation. Some of the nodes will be connected by radiative a n d / o r conductive heat transfer. They can be viewed as forming a double directed graph, one for radiation and one for conduction (Fig. 7.11). The radiant power incident on the ith node is composed of radiation directly emitted from and also reflected from other nodes. In the most
-
Solar or Earth Radiation
-
....
Si (ia,i ih,i)
Radiative Conductive
+
Si q c ~ Si qp,]~
Electric ' ~
(a)
Figure 7.11
Si (Oa,i + Oh,i)
(b)
(a) Thermal inputs to ith node. (b) Directed, double graph for spacecraft thermal calculations.
7.8 L u m p e d Parameter Model of a Spacecraft
293
general case, it includes radiant power, ia, i, per unit area, from ambient temperature sources and similarly, ih, i, from solar radiation, either direct or reflected. The total power per unit area that is directed o u t w a r d from the ith node is (e = emitted power per unit area, i = incident power per unit area) Oa,i - - ea, i + (1
-
(i = 1, 2 . . . . . n)
Ota, i)ia, i
(7.39a)
for ambient temperature radiation and Oh, i - -
JhFh, i(1
-- Ofh, i) %
(1 -
(i = 1, 2, . . . , n)
Olh, i)ih, i
(7.39b)
for solar radiation, where ea, i :
(7.40)
8a, i f f T/4
The quantities jh and Fh, i both refer to the solar radiation. The former is defined as in (7.8). The latter is a purely geometrical factor, the "view factor" as defined below. In addition, as illustrated in the figure, there m a y be added to the node thermal power by conduction, qc, i, defined per unit area of the node. In addition, there m a y be external thermal power, qp, i, per unit area, that m a y be added either by electric dissipation, from combustion of propellant, or from a radioisotope source. (In the case of solar arrays, qp, i is the electrical power delivered by a unit area of the panel and is negative.) The thermal conduction power, qc, i, can be expressed by n qc, i - - E
Fij(Tj - Ti)
(i = 1, 2, . . . , n)
(7.41)
j=l
where F ij is called the c o n d u c t a n c e between nodes i and j. In what follows, steady-state conditions are assumed. Conservation of energy at each node therefore requires that ia, i -- Oa, i -[- ih, i -- Oh, i -1- qc, i + qp, i :
7.8.2
0
(7.42)
View Factors
As a preliminary to the radiative part ofthe calculation, view factors Fij (also called "configuration factors") of pairs (i, j) of nodes are introduced. They are purely geometrical quantities, depending only on the node surfaces Si and Sj and their relative configuration. The nodes, in addition to each having constant temperature, absorptivity, and emissivity, are defined such that each node either has a direct, uninterrupted ray path or none to and from each other node, including itself. The several types of geometric configurations that m a y arise are illustrated in Figs. 7.12 and 7.13. The simple case of two convex surfaces Sr and Sj with direct ray paths between t h e m is illustrated in the first figure. Also shown there is a schematic of a Cassegrain a n t e n n a as an example of the radiative exchange between a concave and a convex surface. The nodes Si a n d Sj are seen to have the required property that all points of Si have direct ray paths with all points of Si itself and with all points of Sj. Another geometrical feature that m a y occur is illustrated by the three plates A, B, C shown in Fig. 7.13a. (B is a thin, two-sided plate; only single
294
C H A P T E ff 7 Spacecraft Thermal Design
Figure 7.12 (a) For derivation of Eq. (7.43). (b) Schematic of Cassegrain antenna as an example of a concave (Si) and of a convex (Sj) surface; Jh -- solar radiation intensity. surfaces of A a n d C p a r t i c i p a t e in the radiative exchange.) To satisfy the definition of nodes, it is n e c e s s a r y h e r e to divide A a n d C e a c h into two nodes. H e n c e a total of six n o d e s w i t h s u r f a c e s $1 to $6 are r e q u i r e d for a c o m p l e t e specification. For instance, $6 h a s a direct ray p a t h (with 0 _< fii, fij < zr/2) to a n d f r o m $1, $2, a n d $4 b u t n o t to a n d f r o m $3, $5, a n d $6. O n e t h e n defines t h e view factor Fij of Si relative to Sj as the ratio of the r a d i a n t p o w e r e m i t t e d a n d reflected by Si a n d i n c i d e n t on Sj to the total r a d i a n t p o w e r e m i t t e d a n d reflected by Si. For diffuse surfaces, as a s s u m e d , t h e o u t w a r d - d i r e c t e d r a d i a n t p o w e r (including e m i s s i o n a n d all reflections), p e r unit area, at t h e angle fl is o 8 On COS ~ w i t h i n the solid angle dzo - 27r sin fi dfi. Therefore
n/2
0 -- 2Jr O~
f COS ~ sin/~ d~ J~-0
~ On
Figure 7.13 (a) Definition ofnodes: panels A and C each consist of two nodes. (b) Illustration of a two-dimensional node pair.
7.8 Lumped Parameter Model of a Spacecraft
295
and
Off - -
7r-locos/~
From this follows, as illustrated in Fig. 7.12a, that the outward directed radiant power from an element of surface dSi of Si that is incident on the surface element dSj of Sj is
dSi:rr-l oi COS i~i dwi
-- 7r-loi
COS
fli COS i~j dSi dSj/Ri2j
where fli and flj are the angles m a d e by the ray with the normals to the surfaces and where Rij is the distance b e t w e e n points on Si and Sj, respectively. Therefore
Fij=
I 1 L L c~176 "-~i i j R~.
(for direct paths b e t w e e n Si a n d Sj, 0 ~ fli, flj < y r / 2 ) (otherwise) (7.43)
0 If i - j, (7.43) can be simplified to
Fie=
l
1 fs
Jr
cos
,
fli COS fl~ dSi
(for direct paths from
R~i
Si to Si, 0 < fli, fl~ < zr/2)
(7.43')
(otherwise)
0
The integration, particularly w h e n involving the fourfold integral in (7.43), tends to be exceedingly c u m b e r s o m e or else trivial (such as for the parallel surfaces considered in Sect. 7.6 where Fij - 1 and Fii -- 0). In spacecraft design practice, the view factors are therefore evaluated numerically by special software. A reciprocity relation
SiFij - SjFji
(i, j - 1, 2 , . . . , n)
(7.44)
holds. It follows immediately from interchanging in (7.43) the indexes i and j. For t w o - d i m e n s i o n a l configurations, such as is the case, for instance, for parallel tubes that are very long c o m p a r e d with their diameters, evaluation by analysis, rather t h a n numerically, is often feasible. Figure 7.13b is representative of such a two-dimensional configuration. The unit vectors n~ and n~ n o r m a l to the surfaces, the angles of incidence fl~ and fi~, a n d the distance vectors Ri*j are all in the transverse plane. It follows easily (with z the lengthwise coordinate) that b e t w e e n dSi at z = 0 a n d dSj at z, --
ni . Rij
=
R~j cos fl'~
,
COS flj .
.
nj . Rij .
.
R~j cos fi~
and Rij - ,/ RT~ + z 2. vt
~
The outward-directed radiant power from the surface element dSi of Si incident on the surface e l e m e n t dSj of Sj is therefore 1
R~j2 cos fl; cos fl~ ~Oi 7r +
dSi dSj
296
C H A P T E R 7 Spacecraft Thermal Design and the radiant power from dSi incident on a strip of width db of Sj is dz --o/ COS fli* COS flj* dSi db f'~ 7r RTj2 -~ (1 + z2/R~2.) 2
The value of the definite integral is fir/2)RTj. Therefore, the radiant power from dSi incident on all of Sj is
1
L cos 13;cos fl~ db
-~Oi dSi
J
R;j
The ratio, therefore, of the outward-directed power from a strip of width dz of Si incident on all of Sj to the total outward-directed power from this strip is 1 L L cosfl~cosfl~ d S j d S i 2Si
i
j
R~j
Therefore the view factor for two-dimensional configurations becomes / 1 fs L cosfl~cosfl~dSjdSi
Fij--
~ii
i
j
R;j
0
(for direct paths between S/and Sj, 0 _< fl*.,t flj zr/2) (otherwise)
(7.45)
If i - j, (7.44) can be simplified to l 1 L cosfl.*t cosfl*' i dS i Fii-
-2 0
,
R~i
(for direct paths from S~to S/, 0 -< ~7, ~*' i < zr/2) (otherwise)
(7.45')
Finally, the view factor Fh, i for solar radiation incident on the ith node is Fh, i --
cos fli dSi
Sii
(for direct paths from the sun)
i
(otherwise)
0
7.8.3
(7.46)
Computational Model
Next, the radiative transfer by emission and reflection among the nodes is considered, first for solar radiation and then for ambient temperature (infrared) radiation. The solar radiant power, Siih, i, incident on the ith node is the sum of directly incident solar radiation and of solar radiation reflected from other spacecraft nodes including itself (for Fii ~ 0). Therefore, making use of the reciprocity relation (7.44), n
s,r
= S, Fh,,jh +
n
Fj, SjOh, j -- S, Fh, gjh + S, j-1
FqOh, j j-1
7.8 Lumped Parameter Model of a Spacecraft
297
a n d substituting for Oh,j from (7.39b) n
ih, i - Fh, gjh + E
Fij[(1 - a h , j)Fh, jjh + (1 --ah, j)ih, j]
(7.47)
j=l
Here the first term on the right represents the effect of direct solar radiation incidence, whereas the s u m a c c o u n t s for reflections of solar radiation, either direct or indirect via still other nodes, that reach the ith n o d e from the various n o d e s (including the ith). Let 1 0
~ij --
when when
i- j i# j
Therefore n
E [ ( ~ i j -- Fij(1 - Olh,j)]ih, j j=l where
I
fh, i = jh Fh, i + ~
n
-
-
(7.48)
fh, i
FijFh, j(1 -C~h,j)
]
(7.49)
We define the square matrix [Mh] and the c o l u m n matrices [ih] a n d [fh] by 1-
Fll
-F21 [Mh] =
(1
-
(1 -
-F12(1 - Olh,2)
Olh,1)
1 --/:22(1
Oth,1)
- 0%2)
....
Fin(1 - Olh,n)
. . . .
F2n(1
--
Olh,n)
.
-Fnl(1
1 - Fnn(1 - ah, n)
- Ofh,1)
(7.50a) [ih] -- [ih, 1, ih,2, "'" ih, n] T
(7.50b)
[fhl -- [fh, 1, fh,2, "'" fh, nl T
(7.5Oc)
so that [Mh] [ih] -- [fh]
Premultiplication with [Mh ~1 results in the final result [ih] - - [ n h ] -
1 [fh]
(7.51)
Having found the incident radiant power, ih, i, per unit area, the corres p o n d i n g outward-directed power, Oh, i, iS found from (7.39b). The d e v e l o p m e n t for the a m b i e n t temperature, infrared radiative exchange a m o n g the n o d e s is similar. Using the reciprocity relation, the radiant power incident on the ith n o d e is n
Siia, i - E j=l
n
FjiSjoa, j - Si E j=l
Fijoa, j
298
C H A P T E R 7 Spacecraft ThermalDesign from which, by substituting for Oa,j from (7.39a) n
ia, i = E
(7.52)
Fij[ea, j + (1 - O t a , j)ia, j]
j=l
so that n
[ ~ i j - F i j ( 1 - Ota,j)]ia, j -- E j=l
(7.53)
Fijea, j
j=l
Here, ea, j can be replaced by means ofthe equation of conservation of energy for the j t h node, because it follows from combining (7.39a), (7.39b), and (7.42) that ea, j -- aa, jia, j + ah, jih, j -- jhFh, j(1 - O~h,j) -[- qc, j --b qp, j
(7.54)
Therefore (7.53) becomes n
Fij]ia, j -- fa, i
(7.55)
Fij[Oth, jih, j -- jhFh, j(1 -- ah, j) + qc, j + qp, j]
(7.56)
E[~ijj=l
where n
fa, i = E j=l
Defining the matrix [Ma] and the column matrices [|a] and [fa] by 1 [Ma] -
Fll
-
- F21
-f~l
-F12
1 -- F22
"~'"
-Fin F2n
. . . .
. . . . . .
1 -
(7.57a)
f~J
[ia] = [ia, 1, ia,2,"" ia, n] T
(7.57b)
[fal -- [fa, 1, fa,2, "'" fa, n] T
(7.57C)
(7.56) can be expressed by [Ma][ia] = [fa]
resulting in the final result for the ambient temperature radiative exchange [ia] -- [ M a ] - l [ f a ]
(7.58)
If conduction a m o n g the nodes can be neglected, that is, qc, i - 0, the algorithm is now complete, because from (7.51) and (7.58), together with the auxiliary equation (7.54), the temperature of each node can be calculated from (7.40). When both radiative and conductive heat transfers are present, an iterative procedure is needed. One such procedure, which in spacecraft thermal calculations is in most cases sufficiently rapidly convergent, is to initially assume qc,~ -- 0, then to calculate an initial set of temperatures that can be used to c o m p u t e a new value for qc,~, and then to repeat the step to obtain increasingly accurate results. (For two-sided surfaces, such as thin sheets or panels, rather than setting the conductance initially to zero, the
7.9 Thermal Control Devices
299
temperatures of the two surfaces can usually be taken to be the same.) The inversions of the matrices need to be calculated only once. In spacecraft, the majority of node pairs will have no direct radiation path connecting the two nodes of the pair. Hence a majority of view factors will be zero, resulting in sparse matrices for [Mh] and [Ma]. This greatly facilitates the numerical inversion and makes it practical to include in the calculation, if needed, several h u n d r e d nodes. Software that is both fast and robust is available for this purpose. To summarize this section: The steady-state thermal transport by nearinfrared and solar radiation and by conduction has been considered. For spacecraft in the vicinity of planets, where planetary thermal and albedo radiation can play a role, these effects can also be included in the analysis, the former by adding the planet as an additional node, the latter by adding terms to the solar radiation. The distinction between the absorptivity O~a for ambient temperature radiation and ah for solar radiation must almost always be made.
7.9
Thermal Control Devices As m e n t i o n e d in Sect. 7.2, passive m e a n s of spacecraft temperature control can be obtained by selecting for the exterior surfaces of the spacecraft coatings or tapes that have the appropriate absorptivity and emissivity. Surfaces that are interior to the spacecraft are also treated in this manner. Black paint is applied to electronic boxes and other devices w h e n electric-tothermal dissipation is significant. An example is the digital module shown in Fig. 7.14. Although some ofthe generated heat flows from the c o m p o n e n t s to
Figure 7.14 Digital module. The EMI (electromagnetic interference) shields are painted black to increase the radiative heat transfer. (From Ref. 8.)
300
C H A P T E R 7 SpacecraftThermal Design the module covers and then to the mounting platform by conduction, additional heat transfer is obtained by radiation from the black paint (applied in this case to the covers that serve to suppress electromagnetic interference). Ideally, thermal control of the spacecraft and of its components would be based on purely passive means, because they are reliable and require no electric power. However, variations in thermal dissipation in on versus off operation, variations in the thermal environment, degradation of the spacecraft exterior surfaces, and a very precisely maintained temperature need for all require active thermal control. Electric heaters, controlled by solid-state devices, are commonly applied for this purpose. Frequently used heating elements are in the form of flexible patches consisting of a conductor in the shape of a pattern that fills the area of the patch, where the conductor is sandwiched between two sheets of Kapton or similar material. Conservative design of the heating elements is very important because otherwise they may fail over long periods of time. Redundancy is obtained by providing two i n d e p e n d e n t circuits, often on the same patch. Spacecraft storage batteries require particularly close thermal control. The requirements can vary greatly, depending on whether the battery is used at a low charge or discharge rate or at high ones. Often, both heating and cooling are needed to meet the varying conditions. Cooling can be obtained by mounting the battery such that one of its surfaces can radiate into space. As indicated in the example shown in Fig. 7.15, this surface is designed as a radiator with low solar absorptivity and high ambient temperature emissivity. In the case of satellites in equatorial or near-equatorial orbits, preferred mounting positions are the north and south panels of the spacecraft, where the solar radiation incidence and its variation are minimal. External heating is provided in this example by patch heaters. Other electric heating elements have the form of standard cartridges that can be inserted and potted into a hole in the component. An example is the cartridge used for preheating the hydrazine catalyst bed of a thruster.
Figure 7.1S Thermal protection of a storage battery on a threeaxis stabilized spacecraft.
7.9
Thermal Control Devices
301
Special forms of heaters have also been developed for wrapping around tubes, for example, for the purpose of preventing hydrazine from freezing and possibly rupturing the tube. The heater current is controlled by means of a temperature-sensing element, usually a thermistor, and a solid-state controller or relay. A small dead band of the controller reduces the temperature swing ofthe component but also increases the number of on-off switch actions, which may reduce the reliability of the circuit. Dead bands of 5 to 10~ are common. Much more precise control, however, is needed for maintaining the alignment and surface figure of some optical systems. Thus, some components of the Hubble Space Telescope are controlled to better than 0.1~ There may be as many as a hundred heating circuits on a spacecraft. They operate autonomously, that is, normally without intervention by the controlling ground station. The temperature readings and controller status are periodically telemetered. There is an advantage to be gained by controlling all heating circuits by the spacecraft's onboard computer, rather than by individual controllers. In case of an unforeseen event or failure, the computer can often be reprogrammed by ground command to circumvent the problem. Even so, autonomous temperature control, as the basic mode of operation, is essential because diagnosing a fault and reprogramming are time consuming. Sending the command may have its own difficulties or, in the case of deep-space probes, may suffer from long communications delays. Louvers represent another type of active control. Typically they are activated by a bimetal spiral spring. Therefore, they require no electric power. Venetian vane-type louvers (Fig. 7.16) consist of rectangular blades that can be rotated into any position between fully open and fully closed. If open, solar radiation is admitted to the surface of the regulated component
Figure 7.16 Bimetal, spring-actuated, rectangular blades ("venetian blind") louvers. Courtesy of LORALSpace & Communications Ltd. By permission.
302
C H A P T E R 7 Spacecraft Thermal Design or spacecraft skin underneath the louvers. If this surface is provided with a high ratio of solar radiation absorptivity to ambient temperature emissivity, control over a wide range of temperatures can be obtained. A ratio of about 6:1 in heat rejection from fully closed to fully open has been achieved. To augment the reliability, each vane is driven separately from the others by its own spring. Other actuators are constructed from bellows, which are sometimes filled with fluid. Louvers with bimetal springs are accurate to better than 10~ More precise control can be obtained by providing a temperature sensor, controller, and a small heater that is in contact with the bimetal spring. This arrangement also has desirable redundancy characteristics: If the heater should fail, the bimetal, mechanical actuator provides a backup, although with less accuracy. Venetian blind louvers can also be used for the temperature control of a component on the side of a spacecraft that is not exposed to solar radiation. In this case, the positions of the louvers merely change the net thermal emission to space by the regulated component. Pinwheel louvers (Fig. 7.17) are similar, but in place of rectangular blades incorporate vanes (typically four) in the form of sectors of a circle that together cover one-half of the area under the pinwheel. They are also rotated by bimetal actuators. Depending on the position of the sectors, they will cover (uncover) areas that have high ratios of solar absorptivity to emissivity and uncover (cover) areas that have a low such ratio. Heat pipes are used on spacecraft and in terrestrial applications for thermal transport and thermal control. Figure 7.18 illustrates a space radiator based on heat pipe technology. A detailed discussion of their theory and design is beyond the scope of this text. For design data on various types of thermal controls on spacecraft, the handbook by Gilmore [8] may be consulted.
Figure 7.17 Bimetal actuated pinwheel assembly for thermal control. (From Gilmore [8].)
Nomenclature
Heat Pipes
Figure 7.18 Space radiator with heat pipes.
Nomenclature
ag C e ~a,i, fh, i
i
J k m n o
q
r
A C1, C2, C3
Fgj Fh, i [Ma], [Mh]
R S T o/
Fij 6i] 8 O" T
r
absorbed radiant power per unit area; a2: thermal diffusivity [Eq. (7.38)] earth albedo coefficient specific heat emitted radiant power per unit area functions defined by Eqs. (7.56) and (7.49) incident radiant power per unit area intensity thermal conductivity mass n u m b e r of nodes outward-directed radiant power per unit area thermal flux per unit area; qc, i: thermal flux added to ith node by conduction; qp, i: thermal flux added to ith node by other sources (electric-to-thermal dissipation, combustion, or radioisotope sources) radial coordinate absorption per unit time universal constants [Eqs. (7.3), (7.4)] view factor for nodes i, j [Eqs. (7.43) to (7.45')] view factor for solar radiation incident on node i matrices defined by Eqs. (7.57a), (7.50a) radius surface absolute temperature absorptivity angle formed by ray with surface normal conductance between nodes i and j [Eq. (7.41)] Kronecker delta emissivity wavelength Stefan-Boltzmann constant = 5.6693 10 -8 W/(m2K 4) nondimensional characteristic time azimuthal angle
303
304
C H A PT E R 7
co, S2
( )a ( )b ( )c ( )g ( )h ( )n ()p
SpacecraftThermal Design solid angles a m b i e n t temperature condition blackbody conduction earth solar; ( )-h eclipse condition normal power
Problems (1) According to Greek mythology, Daedalus fashioned from feathers held by beeswax flying m a c h i n e s for himself and for his son Icarus. But Icarus soared too close to the sun. The wax melted a n d he fell to his death. Beeswax has a melting point of 65~ Its absorptivity for solar radiation is about 0.3; its emissivity at the melting t e m p e r a t u r e is about 0.9. The wax m a y be represented by small spheres, in thermal equilibrium between incident solar radiation and the emission from the spheres. In terms of AUs (astronomical units), to what distance from the sun did Icarus soar before the wax melted? (2) Consider a radioisotope source on a deep-space probe. The source is spherical, with a radius of 0.10 m. The power resulting from the radioactive decay is uniform across the volume and has a density of 0.80 x 106 W / m 3. The thermal conductivity of the material is 5.00 W / ( m K). The outer surface of the sphere is kept at a temperature of 1000 K. A steady-state condition can be assumed. C o m p u t e the t e m p e r a t u r e at the center ofthe sphere. (Solution: 1333 K.) (3) A solar array (a flat panel) of a near-earth spacecraft has solar radiation incident on it at 66.5 ~ before being eclipsed by the earth. The solar-to-electric conversion efficiency is 0.15, the solar radiation reflectivity 0.25. The emissivity of the active side of the array is 0.80, of the back side 0.70. The array is thin e n o u g h so that the temperatures of front and back can be assumed to be the same. The specific heat of the array, per unit area, is 9000 ]/(m 2 K). Earth thermal and albedo radiation can be neglected. The solar radiation intensity is 1350 W / m 2. The preeclipse t e m p e r a t u r e of the array can be a s s u m e d to be equal to the equilibrium t e m p e r a t u r e in the sun. C o m p u t e the array's temperature at 1000 s after the spacecraft has entered the eclipse. (4) Consider two adjoining rectangular flat panels, both of the same width and infinitely long in the third dimension. Express the view factor between the two panels as a function of the angle b e t w e e n them. (5) Find an expression [analogous to Eq. (7.45), which is valid for two-dimensional configurations] to calculate view factors for axisymmetric configurations. (6) Consider a spherical shell. One half of the shell is at temperature T1 and has emissivity el a n d absorptivity Ot 1 on its inside surface. The other half
References
305
is at temperature T2 and has emissivity E2 and absorptivity Or2 on the inside surface. In terms of these quantities, express the net power transferred by radiation from the first to the second hemisphere. (7) Consider a spacecraft consisting of a cylindrical thin shell and a payload inside the shell. The payload m a y be irregularly shaped, although assumed to be convex everywhere. The shell's diameter is 1.00 m, its length 2.50 m. The thickness of the shell is negligible. The cylindrical portion of the spacecraft is covered with solar cells. Solar radiant power at 1350 W / m 2 is incident at a right angle to the cylinder axis. The spacecraft is in the so-called rotisserie mode, spinning fast enough so that the temperature of the shell (including, by conduction, the end caps) can be assumed spacially uniform and constant in time. The solar cells' conversion efficiency is 17%. Their packing factor (which accounts for the gap between cells) is 95%. The reflectivity is 0.20, the emissivity 0.35. The electric power delivered by the solar cells is dissipated in the payload, which has an outer surface of 5.00 m 2 and is assumed to have a uniform temperature. The shell and payload exchange thermal energy by radiation. Conduction is neglected. The inside of the shell and the outside of the payload are painted black, with emissivities and absorptivities of 0.90. Compute the temperature of the shell and of the payload. (Solution: shell temperature = 281 K; payload temperature = 301 K).
References 1.
2. 3. 4. 5. 6. 7. 8.
Horton, T. E., ed., "Spacecraft Radiative Transfer and Temperature Control," Progress in Astronautics and Aeronautics, Vol. 83, American Institute of Aeronautics and Astronautics, Washington, DC, 1982. Siegel, R. and Howell, J. R., "Thermal Radiation Heat Transfer," McGrawHill, New York, 1972. Mills, A. E, "Heat Transfer," Irwin, Boston, 1992. Agrawal, B. N., "Design of Geosynchronous Spacecraft," Prentice Hall, Englewood Cliffs, N], 1986. Pisacane, V. L. and Moore, R. C., eds., "Fundamentals of Space Systems," Oxford University Press, New York, 1994. Hallmann, W. and Ley, W., eds., "Handbuch der Raumfahrttechnik," Carl Hauser Verlag, Munich, 1988. Carslaw, H. S. and Jaeger, J. C., "Conduction of Heat in Solids," 2nd. ed., Clarendon Press, Oxford, 1959. Gilmore, D. G., ed., "Satellite Thermal Control Handbook," Institute of Aeronautics and Astronautics, Washington, DC, 1994.
a This Page Intentionally Left Blank
Physical Constants Used in this Text R e c o m m e n d e d v a l u e s 1986, E. R. C o h e n a n d B. N. Taylor. T h e n u m b e r s in p a r a n t h e s e s r e p r e s e n t t h e d i s p e r s i o n o f t h e last t w o digits o f t h e m e a n value.
Velocity o f light in v a c u u m Universal gravitational constant Planck's c o n s t a n t Boltzmanns's constant Avogadro's n u m b e r U n i v e r s a l gas c o n s t a n t Stefan-Boltzmann constant A t o m i c m a s s c o n s t a n t (12C/12) P r o t o n rest m a s s E l e c t r o n rest m a s s Electronic charge
307
c G h k
= = = =
N R0 a M0 Mp me e
= = = = = = =
2.99792458 108 m s -1 6.67259(85) 10 -11 m 3 kg -1 S - 2 6.6260755(40) 10 -34 J s 1.380658(12) 10-23 J K -1 8.617385(73) 10 -5 eV K -1 6.0221367(36) 1026 kmo1-1 8.314510(70) 103 ] kmo1-1 K -1 5.67051(19) 10 -8 W m -2 K -4 1.6605402(10) 10 -27 kg 1.6726231(10) 10 -27 kg 9.1093897(54) 10 -31 kg 1.60217733(49) 10 -19 C
a This Page Intentionally Left Blank
B Astronomical Constants A s t r o n o m i c a l unit Light year Parsec E p h e m e r i s day Siderial day
AU ly pc dE ds
1.495979 108 k m 9.46054 1012 k m 3.08568 1013 k m 86400 s 86164.09055 s + (0.0015 T) s* dh = 86400 s + (0.0015 T) s* 365.24219 dE 365.25637 dE 365.25 dE 365.2425 dE
M e a n solar day Tropical year Siderial year Julian year Gregorian calendar year
= = = = =
Earth: M e a n equatorial radius Polar radius (based o n spheroid) Mass M e a n density Gravitational p a r a m e t e r Obliquity of ecliptic Gravitational acceleration at e q u a t o r Equatorial rotational velocity Centrifugal acceleration at e q u a t o r I n t e r n a t i o n a l s t a n d a r d gravity Escape velocity at e q u a t o r S u n - e a r t h distance, m i n i m u m maximum
6378.140 kin 6356.755 kin 5.976 1024 kg 5.517 g c m -3 3.986013 105 k m 3 s -2 23~ T* 9.8142 m s -2 0.4651 k m s -I 0.0339 m s -2 9.80665 m s -2 l l . 1 9 k m s -I 1.4710 lO 8 k m 1.5210 lO 8 krn
Moon: M e a n radius Mass M e a n density Gravitational p a r a m e t e r M e a n orbital inclination to ecliptic M e a n equatorial inclination to ecliptic Rotation p e r i o d Gravitational acceleration at surface Escape velocity at surface E a r t h - m o o n distance, m i n i m u m maximum
1738.2 k m 7.350 1022 kg 3.341 g c m -3 4.90265 103 k m 3 s -2 5~ 1~ 27.32166 dE 1.62 m s -2 2.38 k m s -1 356 400 k m 406 700 k m
*T is in centuries from the year 1900.
309
310
Appendix B Astronomical Constants Sun:
6.9599 105 k m 1.990 1030 kg 1.409 g c m -3 1.32712 1011 km 3 s -2 7o15 ' 25.38 days 273.97 m s -2 0.0057 m s -2 3.826 1026 W 1360 W m -2 5760 K
Mean radius Mass M e a n density Gravitational p a r a m e t e r Inclination of e q u a t o r to ecliptic Rotation period (at 17 ~ latitude) Gravitational acceleration at surface Centrifugal acceleration at surface Radiation emitted Solar c o n s t a n t Effective blackbody t e m p e r a t u r e
Planets: Mercury Max. d i s t a n c e from s u n (AU) Min. d i s t a n c e from s u n (AU) Siderial p e r i o d (tropical years) M e a n orbital velocity (km s- I) Equatorial radius (km) Mass ( E a r t h - 1) M e a n density (g c m -3)
Venus
Earth
Mars
Jupiter
Saturn
Uranus
Neptune
Pluto
0.467
0.728
1.017
1.666
5.452
10.081
19.997
30.340
48.94
0.307
0.718
0.983
1.381
4.953
9.015
18.272
29.682
29.64
0.2408
0.6152
1.000
1.8809
11.861
29.50
83.70
164.79
246.28
47.87
35.02
29.78
24.13
13.06
9.64
6.81
5.44
4.75
2432
6052
6378
3402
71490
60,000
25,400
24,765
3200
0.0558 5.53
0.8150 5.25
1.000 5.52
0.1074 3.93
317.89 1.36
95.179 0.71
14.629 1.31
17.222 1.66
0.111 4.83
8.87
9.81
3.73
25.7
10.8
8.95
11.0
4.3
0.0
0.034
0.017
2.23
1.75
0.66
0.31
0.0
243 d
23h56m4 s
24h37m23 s 9h50 m
10h49 m
16h07 m
3.76 Equatorial gravitational a c c e l e r a t i o n ( m s -2) Equatorial centri0.0 fugal a c c e l e r a t i o n ( m s -2) 59 d Siderial r o t a t i o n period 0o Equator's inclination to orbit Virtually Atmosphere, main none components
retrograde 177 ~
23~
Carbon dioxide
Nitrogen oxygen
10 h14 m
,
23~
'
Carbon dioxide
3~5 ,
26~
'
Hydrogen helium
Hydrogen helium
97~
'
Helium hydrogen methane
6d9 h retrograde
retrograde 28~
'
Hydrogen helium methane
None detected
Heliocentric Osculating Elements of the Planets for 1996 The elements listed here are referred to the m e a n ecliptic and equinox at the Julian date 2000.0. First row u n d e r each planet is for the Julian date 2 450 120.5, the second row for 2 450 320.5. Values given for Earth are actually for the E a r t h - m o o n barycenter. Taken from The Astronomical Almanac for the Year 1996, U.S. G o v e r n m e n t Printing Office, Washington, DC and Her Majesty's Stationery Office, London.
Appendix B Astronomical Constants
Ecliptic longitude Orbit inclination Ascending node Perihelion (deg)
311
Meg)
(deft,)
Mean distance (AU)
Mean angular velocity (deg/ Eccentricity sidereal day)
48.3362 48.3356
77.4503 77.4537
0.3870985 0.3870975
0.2056337 0.2056430
4.092342 4.092358
131.470 131.814
0.7233257 0.7233235
0.0067574 0.0067974
1.602152 1.602159
102.8913 102.8837
1.0000082 1.0000108
0.0167549 0.0167206
0.9855970 0.9855931
49.5709 49.5705
336.0217 336.0365
1.5236253 1.5237131
0.0932886 0.0933046
0.5240671 0.5240218
100.4708 100.4707
15.7513 15.7225
5.202289 5.202427
0.0484146 0.0484362
0.08310355 0.08310024
113.6415 113.6365
90.5489 89.8160
9.555375 9.561943
0.0523713 0.0525151
0.03337295 0.03333857
74.0909 74.0971
176.6862 176.6634
19.29767 19.30425
0.0444402 0.0437359
0.01162669 0.01162075
131.7789 131.7851
2.445 2.167
30.26605 30.27750
0.0084404 0.0092486
0.005919454 0.005916098
110.3902 110.3949
224.7317 224.8114
39.77030 39.71320
0.2538350 0.2526847
0.003929758 0.003938236
Mercury 7.00519 7.00513
Venus 3.39480 3.39477
76.6916 76.6900
Earth 0.00051 0.00047
345.4 346.4
Mars 1.84997 1.84991
Iupiter 1.30461 1.30460
Saturn 2.48534 2.48525
Uranus 0.77337 0.77345
Neptune 1.76969 1.76902
Pluto 17.11942 17.11829
The Julian Date of an event is the n u m b e r of days, and fractions thereof, elapsed after an origin of time set at 4713 B.C., January 1, Greenwich noon. In this reckoning, there is a leap year every 4 years (hence the designation "Julian"). A Julian day starts at Greenwich m e a n n o o n until the next noon. For example, the Julian date for January 1 in the year AD 2000 in the m o d e r n calendar, at 18 hours Universal Time is 2 451 545.25. The values given in the table are for the osculating elements. These are p a r a m eters that specify the i n s t a n t a n e o u s position a n d velocity of a celestial b o d y that the body would follow if perturbations were to cease instantaneously.
a This Page Intentionally Left Blank
C (a) E a r t h A t m o s p h e r e a b o v e 100 k m A l t i t u d e . 1976 U.S. S t a n d a r d A t m o s p h e r e
Altitude (km)
Temperature (K)
Pressure (N/m 2)
Density (kg/m 3)
100
195
3.20 10 -2
5.60 10 -7
110
240
7.10 10 -3
9.71 10 -8 2.22 10 -8
120
360
2.54 10 -3
130
469
1.25 10 -3
8.15 10 -9
140
560
7.20 10 -4
3.83 10 -9
150
634
4.54 10 -4
2.08 10 -9
160
696
3.04 10 -4
1.23 10 -9
170
748
2.12 10 -4
7.81 10 - l ~
180
790
1.53 10 -4
5.19 10 -1~
190
825
1.13 10 -4
3.58 10 -1~
200
855
8.47 10 -5
2.54 10 -1~
210
879
6.48 10 -5
1.85 10 -1~
220
899
5.01 10 -5
1.37 10 -1~
230
916
3.93 10 -5
1.03 10 -1~
240
930
3.11 10 -5
7.86 10 -11
250
941
2.48 10 -5
6.07 10 -11
260
951
1.99 10 -5
4.74 10 -11
270
959
1.61 10 -5
3.74 10 -11
280
966
1.31 10 -5
2.97 10 -11
290
971
1.07 10 -5
2.38 10 -11
300
976
8.77 10 -6
1.92 10 -11
310
980
7.23 10 -6
1.55 10-11
320
983
5.98 10 -6
1.26 10 -11
330
986
4.96 10 -6
1.03 10 -11
340
988
4.13 10 -6
8.50 10-12
350
990
3.45 10 -6
7.01 10 -12
360
992
2.89 10 -6
5.80 10-12
370
993
2.43 10 -6
4.82 10 -12
380
994
2.04 10 -8
4.01 10 -12
390
995
1.72 10 -s
3.35 1 0 -12
400
996
1.45 10 -6
2.80 10 -12
313
314
Appendix C (b) Mars A t m o s p h e r e : N o m i n a l M o d e l s of Daily M e a n at M i d l a t i t u d e
Northern s u m m e r z (km)
T (I0
Southern s u m m e r
P (N/m2)
0 (kg/m3)
P (N/m2)
0 (kg/m3)
636
1.56 10 -2
730
1.78 10 -2
0
214
2
214
530
1.30
608
1.49
4
213
441
1.08
507
1.24
6
212
368
9.07 10 -3
423
1.04
8
209
306
7.65
351
8.78 10 -3 7.42
10
205
254
6.47
291
12
201
210
5.45
241
6.25
14
198
173
4.57
198
5.24
16
195
142
3.81
163
4.37
18
191
116
3.17
133
3.64
20
188
94.7
2.63
109
3.02
22
185
77.0
2.18
88.4
2.50
24
183
62.5
1.79
71.7
2.05
26
180
50.6
1.47
58.1
1.69
28 30
178 175
40.8 32.8
1.20 9.84 10 -4
46.8 37.6
1.38 1.13
32
173
26.3
7.98
30.2
9.16 10 -4
34
170
21.1
6.48
24.2
7.44
36
168
16.8
5.24
19.3
6.01
38
165
13.3
4.23
15.3
4.86
40
162
10.6
3.40
12.2
3.90
42
160
8.33
2.72
9.56
3.12
44
158
6.56
2.17
7.53
2.49
46
156
5.15
1.73
5.91
1.99
48
154
4.03
1.37
4.63
1.57
50 60
152 144
3.15 8.79 10 -1
1.08 3.19 10 -5
3.62 1.01
1.24 3.66 10 -5
70
140
2.33
8.75 10 -6
2.67 10 -1
1.00
80
139
6.09 10 -2
2.29
6.99 10 -2
2.63 10 -6
90
139
1.60
6.03 10 -7
1.84
6.92 10 -7
100
139
4.24 10 -3
1.60
4.87 10 -3
1.84
Pressures are inferred f r o m post-Viking Mission p a r a c h u t e descents. Altitude z is f r o m h e i g h t a b o v e r e f e r e n c e ellipsoid.
R e p r i n t e d f r o m A. Seiff, in
'Tkdvances of Space Research," Vol. 2, No. 2. Copyright 9 1982, w i t h p e r m i s s i o n from Elsevier Science.
Appendix C Mars Atmosphere: Principal Composition Gas
Mole fraction
CO2
0.9555 --1-0.0065 0.027 4- 0.003 0.016 -4- 0.003 0.0015 4- 0.0005
N2 At" 02
Mean molecular weight = 43.49 4- 0.07; gas constant = 191.18 4- 0.29 ]/kg K.
315
a This Page Intentionally Left Blank
D Properties of Selected Rocket Propellants Sources: Chemical Propulsion Information Agency (CPIA) manuals, Johns Hopkins University, Baltimore, MD (continuing database); Sutton, G. P. and Ross, D. M., Rocket Propulsion Elements, 5th ed., John Wiley & Sons, New York, 1986.
(a) Liquid, Noncryogenic Propellants Propellant
Chemical symbol
Use
Molecular weight
Freezing point (~
Boiling point at 1 atm (~
Vapor pressure (N/cm 2)
Density (g/cm 2)
Nitrogen tetroxide
N204
Oxidizer
92.0
-12
21
77 at 70 ~C
1.44 at 20 ~C
IFRNA (inhibited) fuming red nitric acid)
82% HNO3 15% NO2 2% H20 1% HF BrF5
Oxidizer, coolant
55.9
-49
66
11.9 at 70~
1.57 at 20 ~C
Oxidizer, coolant
174.9
-62
40
28.3 at 70 ~C
2.48 at 20 ~C
Tetranitro methane
C(NO2)4
Oxidizer
196.0
14
126
1.64 at 75 ~C
at 20 ~C
Chlorine trifluoride
C1F3
Oxidizer
92.5
-76
11.8
55 at 60 ~C
1.83 at 20~
RP-1 (rocket propellant) UDMH (unsym. dimethylhydrazine) 92.5% ethyl alcohol
Hydrocarbons (CH3)2NNH2
Fuel, coolant Fuel, coolant
165 to 195
- 4 4 to - 5 3
172 to 264
60.8
-58
63
0.23 at 70~ 12.1 at 70 ~C
.80 to .82 at 20 ~C 0.789 at 20~
C2H5 OH
Fuel, coolant
41.2
-123
78
8.95 at 70 ~C
0.81 at 15~
JP-4 (jet propulsion fuel) Pentaborane
Hydrocarbons
Fuel, coolant
128
-60
130 to 240
4.95 at 70 ~C
0.75 to 0.82
BaH9
Fuel,
63.2
-47
60
13.1 at 70~
0.61 at 20 ~C
Propyl nitrate Trimethyl amine 95% hydrogen peroxide
C5H7NOa
Fuel, coolant Fuel
105.1
-91
111
59.1
-117
2.8
32.6
-5.6
146
2.55 at 70 ~C} 74.5 at 70 ~C 0.035 at 25~
1.06 at 20~ 0.603 at 20~ 1.414 at 25~
32.1
1.4
113
1.93 at 70~
1.01 at 20~
Bromine pentafluoride
Hydrazine
(CHa)aN H2 02
N2H4
Monopropl. oxidizer, coolant Monopropl., coolant
317
1.64
Compatible materials Stainless steel; A1, Ni alloys; Teflon Stainless steel; AI alloys; polyethylene A1 alloys; 18-8 stainless steel; Teflon A1 alloys; mild steel; Teflon AI alloys; 18-8 stainless steel; Ni alloys; Teflon AI, steel, Ni alloys; Teflon; neoprene A1 alloys; stainless steel; Teflon A1, steel, Ni alloys; Teflon, polyethylene A1, steel, Ni alloys; neoprene; Teflon AI alloys; steel; copper; Teflon; viton A1 alloys; stainless steel; Teflon AI alloys; steel; copper; Teflon AI alloys; stainless steel; Teflon A1 alloys; 304, 307 stainless steel; Teflon
318
Properties of Selected Rocket Propellants
Appendix D
(b) C r y o g e n i c P r o p e l l a n t s
Density Freezing Boiling Critical Critical at boiling Molecular point point pressure temp. point (N/cm 2) (~ (g/cm 3) weight (~ (~
Chemical
Propellant
symbol
Use
Liquid oxygen
02
Oxidizer
32.0
Oxygen difluoride
OF2
Oxidizer
54.0
Liquid fuorine
F2
Oxidizer
38.0
Liquid hydrogen Ammonia
H2
Fuel, coolant Fuel, coolant
NH3
-219
2.016 17.03
-183
508
-119
1.142
-184
496
-58
1.521
-220
-188
557
-129
1.509
-259
-253
127
-240
0.071
- 78
- 33
1092
132
0.683
Compatible materials Stainless steel; AI, Ni alloys; copper; Teflon AI alloys; 300 series stainless steel; Ni alloys; brass AI alloys; 300 series stainless steel; Ni alloys;brass Stainless steel; Ni alloys AI alloys; steel; Teflon
(c) Solid P r o p e l l a n t Oxidizers
Oxidizer
Chemical symbol
Molecular weight
Ammonium perchlorate
NH4C104
Potassium perchlorate
KC104
Lithium perchlorate Sodium perchlorate
Density (g/cm 3)
Oxygen content by mass
117.5
1.95
34%
138.6
2.52
46%
LiC104
106.4
2.43
60%
NaC104
122.4
2.02
52% 39%
Ammonium nitrate
NH4NO3
80.0
1.73
Lithium nitrate
LiNO3
68.9
2.38
58%
Sodium nitrate
NaNO3
89.0
2.26
47%
Nitronium perchlorate
NO2C104
145.5
2.20
66%
(d) Solid P r o p e l l a n t Fuels
Melting point
Chemical symbol
Molecular weight
Aluminum
Al
26.98
659
2.70
PU-AP-AI: 265 s
Beryllium
Be
9.01
1277
Boron
B
10.81
2304
1.84 2.30 (1)
PU-AP-B: 255 s
Aluminum
AlH3
30.0
decomposes
1.42
PU-AP-AIH3" 280 s
Bell2
11.03
decomposes
0.65 (2)
PU-AP-BeH2:310 s
Fuel
(~
Density (g/cm 3)
Typical Isp for listed propellant combination, for expansion from 690 N/cm 2 to sea level pressure
PU-AP-Be: 280 s
hydride Beryllium hydride PU, p o l y u r e t h a n e ; AP, a m m o n i u m p e r c h l o r a t e . (1) Crystalline; (2) c o m p a c t e d .
E Thermal Properties of Selected Spacecraft Materials Unless otherwise noted, the values are for 300 K
(kg]m 3)
Thermal conduct, W/(m I0
Specific heat I/(kg K)
2770
177
875
at 600 K
65
473
at 400 K
163
787
at 200 K
186
925
at 100 K
186
1042
179 43.6
1862 1046
25.2
Density Material A l u m i n u m 2014-T6 (4.5% Cu,
Thermal expansion (10-6/K) 73.0
1.5% Mg, 0.6% Mn)
B e r y l l i u m alloy (extrusion) M a g n e s i u m (extrusion, AZ31B)
1850 1770
T i t a n i u m 6Al-4V
4430
C o n s t a n t a n (55% Cu, 45% Ni)
8920
at 200 K at 100 K Iridium
22500
7.4
11.5
502
8.8
23
384
6.71
17
237
19
362
147
130
at 400 K
144
133
at 800 K
132
142
at 1500 K
111
50.3
172
C a r b o n steel (Mn < 1%, Si < 0.1%)
7854
60.5
434
17.7
C a r b o n steel (1% < M n < 1.65%,
8131
41.0
434
11.6
0.1% < Si < 0.6%) C h r o m i u m - v a n a d i u m steel
7836
48.9
443
14.1
8055
(0.2% C, 1.02% Cr, 0.15% V) Stainless steel AIS1302
15.1
480
at 1000 K
17.3
512
at 800 K
20.0
559
at 600 K
22.8
585
at 400 K
25.4
606
14.9
477
at 1500 K
9.2
272
at 1200 K
12.6
402
at 800 K
16.6
515
at 400 K
22.6
582
Stainless steel AIS1304
7900
3.91
3.95
(Continued) 319
320
Appendix E
Thermal Properties of Selected Spacecraft Materials
(Continued) i
Material at 200 K at 100 K N i c h r o m e (80% Ni, 20% Cr) at 800 K at 600 K at 400 K Inconel X-750 at 1500 K at 1000 K at 600 K at 400 K Invar 36, a n n e a l e d Tantalum at 2500 K at 1500 K at 1000 K at 600 K Tungsten at 600 K at 1000 K at 1500 K at 2500 K Boron fiber epoxy, 30% vol. at 400 K parallel perpendicular
Density (kg/m 3)
8400
8510
8080 16600
19300
Thermal conduct. W/(m K)
Specific heat I/(kg K)
28.0
640
31.7 12 14 16 21 11.7 13.5 17.0 24.0
682 420 480 525 545 439 473 510 626
33.0 13.5 57.5 58.6 60.2 62.2 65.6 174 137 118 107 95
514 140 146 152 160 189 132 142 148 157 176
3.1
1.26 24.7
68.3
364 2.10 0.37 757 2.23 0.49
at 200 K parallel perpendicular
2.29 0.59
at 300 K parallel perpendicular
3.4
2080
at 300 K parallel perpendicular
at 100 K parallel perpendicular Graphite fiber epoxy, 25% vol. at 400 K parallel perpendicular
Thermal expansion (10-elK)
1122
1431 2.28 0.60 1400 337 5.7 0.46 642 8.7 0.68
(Continued)
Appendix E Thermal Properties of Selected Spacecraft Materials
321
(Continued) Thermal
Material at 200 K parallel perpendicular at 100 K parallel perpendicular
Density
conduct.
(kg/m 3)
W/(m K)
Specific heat J/(kg K) 935
ii.I 0.87 1216 13.0 I.I
Thermal expansion (10-6/K)
a This Page Intentionally Left Blank
F Absorption and Emission Coefficients of Spacecraft Materials Unless o t h e r w i s e noted, the values are for surfaces n o t yet d e g r a d e d by the space e n v i r o n m e n t . A d a p t e d f r o m Gilmore, D. G., ed., "Satellite T h e r m a l Control H a n d b o o k , " A m e r i c a n Institute of A e r o n a u t i c s a n d Astronautics, Washington, DC, 1994.
OCa ~ Ca
solar
ambient temperature
.96 .95 .93 .92 .96 .97 .95 .97 .97 .94
.88 .87 .87 .87 .88 .73 .75 .91 .84 .90
.24 .20 .17 .15 .06 .39
.90 .90 .92 .92 .88 .87
.53-.67 .73 .44 .48
.82-.87 .86 .56 .82
.16 .15
.03 .05
~h
Material Black coatings and plastics Carbon black paint NS-7 Black polyurethane paint Z306, 3 mil thick after 3 years GEO after 5 years GEO CATALAC black paint EBANOL carbon black PALADIN black lacquer 3M Black Velvet paint after 2.5 years GEO TEDLAR black plastic White coatings and plastics CATAIAC white paint Titanium oxide with methyl silicate Titanium oxide with potassium silicate Zinc oxide with sodium silicate Barium sulphate with polyvinyl alcohol TEDLAR white plastic Anodized aluminum Black or blue anodized aluminum Brown anodized aluminum Chromic anodized aluminum Gold anodized aluminum Metals Buffed aluminum Polished aluminum
(Continued) 323
324
Appendix F Absorption and Emission Coefficients of Spacecraft Materials
(Continued) O~a ~
solar .13 .42 .40 .52 .40 .44 .30
.30 .11 .05 .10 .05 .03 .05
.23
.03
.72
.89
.93
.85
.07 .11 .12 .35 .14 .04
.76 .76 .03 .53 .28 .10
r
Material Heavily oxidized aluminum Polished stainless steel Stainless steel 1 mil foil Inconel foil Tantalum foil Polished tungsten Polished gold same after 5 years GEO Electroplated gold Composites Fiberglass epoxy same after 5 years GEO Graphite epoxy same after 5 years GEO Miscellaneous Indium oxide optical solar reflector after 5 years GEO Aluminized Kapton, first surface Aluminized Kapton 0.5 mil, second surface Mylar film 0.15 mil with aluminum backing Aluminum tape
•a
ambient temperature
Index absides, 193 absorption ambient temperature radiation, 278 blackbody; 270 solar radiation, 275 technical surfaces, 275 acceleration, 7 at equator, 2, 28 residual, 2 transformation equation, 8 accelerometers, 239 accommodation coefficient, 44 actuators, s e e a l s o thrusters, 240 aerobraking, 37 aerodynamic forces and moments, 37 aerospike, 115, 121 albedo, 285 ammonium nitrate, 171 ammonium perchlorate, 171 anomaly eccentric, 65 hyperbolic, 66 mean, 65, 69 true, 61, 70, 207 apoapsis, apogee, aphelion, 61,190, 194, 203 ascending node, 69 ascent, 37 atmosphere atomic oxygen, 280 density, 42, 52 effects, 37, 39, 40, 73 entry, s e e reentry ionosphere, 73 Mars, 120 rocket motor operation, 117 scale height, 52 skipping, 54 temperature, 42, 52 upper atmosphere wind, 39 atomic clocks, 15 oxygen, 280
325
time (TAI), 15, 17 attitude control, 215 to 265 azimuth, 4 back pressure, 117 barycenter, 6, 63 batteries, 300 bell nozzle, 115, 121 Bernoulli's equation, 152 beryllium, 171 bipropellants, 140, 167 blackbody, s e e thermal emission, thermal absorption body cone, 227 Boltzmann, 272, 273 boosters, s e e a l s o launch vehicles, 101 Bryson, A. E., 218 buffeting, 40 burn rate, 172 capillary feeds, 168 capture, 89 catalyst bed, 142, 300 cavitation, 151 colatitude, 4 collisions, 11 combustion chamber, 100 instabilities, 156 composite propellants, 170 contamination, 123 Coriolis force, 9 cryogens, 140 D'Alembert, 66 Dalton's law, 128 debris, 280 declination, 6, 70 degree of reaction, 129, 132 De Laval, 115 desaturation, 246, 250 despun platform, 216 diaphragms, 146
326
Index
docking, 24, 205, 208 Doppler shift, 74, 240 double base propellants, 170 drag, 37 drag free satellite, 46 dual spin, 258 earth albedo, 285 diurnal rotation, 17, 65 gravity, 26 reference ellipsoid, 27 thermal emission, 284 eccentricity, 62, 68 ecliptic, 3, 69, 284 latitude, longitude, 6 obliquity, 4, 284 eigenvectors, 219 Einstein, 92 electric propulsion, 99 thrusters, 245 elevation, 4 emission, thermal blackbody, 270 earth, 284 solar, 273 technical surfaces, 275 epoch, 14 equatorial plane, 69 celestial, 3, 69, 284 equilibrium constant, 131 equinox, 3, 69, 284 Euler angles, 12, 226, 230, 239 Euler's equation for rigid bodies, 224, 263 for time dependent inertia, 223 Evans form, 256 exit plane, 100, 104 expansion-deflection nozzle, 115, 121 extendable nozzle, 115, 122 f and g series film cooling, 149 first point of Aries, 3, 284 flow rate control, 162 flyby, 85 forces, see also gravity, thrust aerodynamic, 37 buffeting, 40 Coriolis, 9 drag, lift, 37, 211 gravitational, 21 inertial, 9
Fourier, 289 free molecule flow, 40 heat transfer, 47 pressure, 45 shear, 45 free radicals, 97 frozen equilibrium, 126, 131,135 fuels, 141, 171 hydrocarbons, 141 hydrogen slush, 142 liquid hydrogen, 141 unsymmetric dimethylhydrazine, 141 Galilean invariance, 2 gas-dynamic shocks, 117, 119, 122 gas generators, 147, 149 gas turbines, 147 Gauss, 23, 27 geocentric reference, 5 geoid, 27 geostationary orbit, 77 Gibbs, 129, 131 gimbals, 160, 236 Global Positioning System, 16, 64 Goddard, 147 grain recession, 172 gravitational parameter, 22, 60 gravity, 21 assist, 85 gradient, 24, 216, 262 loss, 35, 109 turn, 209 Greenwich mean solar time (GMST), 16 meridian, 4 grey surfaces, 278 ground track, 4 gyroscopes, 236 integrating, 239 rate, 238 heat pipes, 302 heat transfer, 47, 49, 127, 146, 149, 157, 269 heaters, 300 heliocentric reference, 6, 87 Hohmann transfers, 192, 202, 206 bielliptic, 196 modified, 195 semitangential, 196 horizon, 71,232 sensors, 232 Huang, D. H., 99
Index
Huzel, D. K., 99 hybrid motors, 103, 175 propellants, 139 hydrazine, 140, 142, 165 hydrocarbons, 141 hydrogen, 141 para- and ortho-, 97 peroxide, 142 slush, 142 hyperbolic anomaly, 66 hypergolic propellants, 140 igniter, 101, 156 impact parameter, 87 impulsive thrust, 188 inclination, 3, 68 inertia moments of, 218 principal moments of, 221 tensor, 25 inertial forces, 9 measurement units, 239 space, 2 inertially symmetric body, 221 stability, 227 inhibited red fuming nitric acid, 138 injector plate, 99, 155 insertion errors, 199 eccentricity, 202 inclination, 199 semimajor axis, 201 inverse square field, 23 ionosphere, 73 jet damping, 252 Kepler, 60 equation, 65 laws, 62 kinetic energy of rotation, 227 Kirchhoff's law, 276 Knudsen number, 41 Lagrange, 107 Lambert's law, 272, 284 theorem, 184 Laplace, 27, 85 invariable plane, 7 latitude, 4, 6 launch vehicles, see also boosters, 99, 104, 108, 117, 143, 209
vibrations, 157 launch windows, 197 law of mass action, 130, 138 leap second, 16 liquid propellants, 99, 138, 140 longitude, 4, 6 louvers, 301 low earth orbits, 76 lumped parameter model, 290 magnetic levitation, 248 torque, 216 Marman clamps, 104 mass action law, 130, 138 fractions, 129 mean free path, 41 micrometeoroids, 280 microwave communications, 77 minimum energy paths, 181 mole fractions, 128 Molniya, 63 moments of inertia, 218, 221 m o m e n t u m wheels, 249 monomethylhydrazine, 138 monopropellants, 140 hydrazine, 142, 165 hydrogen peroxide, 142 motor case, 175 multilayer insulation, 286 Newton, 2, 21, 23, 33, 62 nitric acid, 141 nitrogen tetroxide, 140 north-south correction, 205 nozzle, 100, 117, 149 aerospike, 115, 121 bell, 115, 121 divergence angle, 125 exit plane, 100, 104 expansion-deflection, 115, 121 extendible, 115, 122 over-expansion, 119, 123 throat, 100, 104, 113 under-expansion, 119, 123 nutation, 229 angle, 13, 226 damping, 255, 258 oblate vehicles, 225, 255 orbits angular momentum, 60
327
328
Index
Orbits (contd.) apoapsis, 61 ascending node, 69 circular, 75 eccentricity, 62 elements, 68 energy, 60 geostationary; 77 inclination, 68 injection errors, 199 Kepler, 60 low earth, 76 on orbit drift, 196, 202 periapsis, 61 period, 64 right ascension of the ascending node, 69 semimajor axis, 63 sun synchronous, 80 oxidizers, 140, 171 liquid oxygen, 140 nitric acid, 141 nitrogen tetroxide, 140 red fuming nitric acid, 141 white fuming nitric acid, 141 paints, 279 panels, 286 partial pressure, 128 periapsis, perigee, perihelion, 61, 69, 93, 190, 194, 203 Planck's radiation law, 272 plane change, 188 planets flyby, 85 missions to, 81, 187 planetocentric reference, 4, 86 pogo oscillation, 158 Prandtl-Meyer function, 124 precession, 226 angle, 12 pressure thrust, 32 principal axes, 218 prolate vehicles, 225, 256 propellants, see also fuels, oxidizers binders, 171 bipropellants, 140 composite, 170 cryogenic, 140 double base, 170 feed systems, 147 flow rate control, 162 hybrid, 139
hypergolic, 140 monopropellants, 142 pumps, 148, 151 solid, 139 tanks, 143 range rate, 73 reaction control motors, see thrusters reaction wheels, 245 desaturation of, 246, 250 magnetically levitated, 248 reentry, 49 reference frames accelerated, 7 barycentric, 6, 63 geocentric, 5 heliocentric, 6, 87 pitch, roll, yaw, 230, 235 topocentric, 4 reflection of thermal radiation by technical surfaces, 275 relativistic effects, 91 rendez-vous maneuver, 205 right ascension, 6, 69 rocket motors, 99, 111,135 see also thrust, nozzle case, 175 characteristic velocity, 106 combustion instabilities, 156 controls, 162 gimbals, 160 hybrid, 103, 175 idealized model, 111 igniter, 101, 156 injector, 99, 155 liquid propellant, 99 operation in the atmosphere, 117 regenerative cooling, 149 shut-down, 155, 164 slag, 32 solid propellant, 101, 170 start-up, 155, 164 throttling, 155 thrust chamber, 100, 155 thrust vector control, 160, 166 Ross, D. M., 99 satellites, see spacecraft Schmitt trigger, 242 second leap, 16 Syst~me International (SI), 15 second surface mirrors, 279 sensors, 230
Index
horizon, 232 star, 230 sun, 231 shifting equilibrium, 126, 131, 137 sidereal time, 17 sodium perchlorate, 171 solar constant, 274 photosphere, 274 radiation pressure, 21, 47, 216 sailing, 21, 47 thermal radiation, 273 time, 16 ultraviolet, 280 solstice, 284 sounding rocket, 34 space cone, 227 spacecraft actuators, 240 despun platform, 216 dual spin, 258 gravity gradient stabilized, 262 main motors, 165 observations, 73 position as a function of time, 65 sensors, 230 spin stabilized, 216, 251 three-axis stabilized, 217, 232 thrusters, 165, 229 tumbling, 14 visibility above horizon, 71 Space Shuttle, 46, 161,174 specific impulse, 33, 99, 137, 139 sphere of influence, 84, 86 spin, 227 squibs, 174 stages, 103, 191 residual mass, 105 separation, 104 Stefan-Boltzmann law, 273 stoichiometric coefficients, 130 strap-on motors, see boosters Summerfield criterion, 123 Sutton, G. E, 99 synoptic period, 207 tesseral harmonics, 29 tether, 24 thermal absorption, 269 blankets, 286
329
conduction, 289 control devices, 299 diffusivity, 289 emission, 269 thrust, 31, 99, 109 control, 162 ideal, 116 in the atmosphere, 117 impulsive, 188 pressure thrust, 32, 116 shut-down, 155, 164 start-up, 155, 164 vacuum thrust, 99, 119 vector control, 160 velocity thrust, 32, 116 thrusters, 127, 217, 241 control, 242 dead band, 242 electric, 245 time atomic (TAI), 15, 17 derivative in different reference systems, 7 Greenwich mean solar time (GMST), 16 sidereal, 17 solar, 16 standard, 16 universal (UT), 16 torques aerodynamic, 37, 216 gravitational, 24, 216 magnetic, 216 solar radiation, 216 Tsiolkovsky, 34 turbo-pumps, 147 ullage, 144 universal gravitational constant, 22 unsymmetric dimethylhydrazine, 138 Van Allen belts, 79 Vertregt, M., 107 vibration modes, 126 view factors, 293 wake flow, 121 Wien's displacement law, 273 yo-yo mechanism, 10 zonal harmonics, 29, 82
a This Page Intentionally Left Blank